Proteins at Solid-Liquid Interfaces

PRINCIPLES AND PRACTICE Philippe Déjardin (Ed.) Proteins at Solid–Liquid Interfaces With 165 Figures and 32 Tables ...

Author: Philippe Déjardin

75 downloads 1297 Views 6MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

PRINCIPLES AND PRACTICE

Philippe Déjardin (Ed.)

Proteins at Solid–Liquid Interfaces With 165 Figures and 32 Tables

123

Dr. Philippe Déjardin Directeur de Recherche CNRS IEM – Université Montpellier 2 CC 047 2 Place Eugène Bataillon 34095 Montpellier Cedex 5 France e-mail: [email protected]

Library of Congress Control Number: 2006927284

ISBN-10 3-540-32657-X Springer Berlin Heidelberg New York ISBN-13 978-3-540-32657-1 Springer Berlin Heidelberg New York

This work is subject to copyright. All rights reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable for prosecution under the German Copyright Law. Springer is a part of Springer Science+Business Media springer.com © Springer-Verlag Berlin Heidelberg 2006 The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Editor: Dr. Sabine Schreck, Heidelberg, Germany Desk Editor: Dr. Jutta Lindenborn, Heidelberg, Germany Cover design: design&production, Heidelberg, Germany Typesetting and production: LE-TEX Jelonek, Schmidt & Vöckler GbR, Leipzig, Germany 31/3100-YL - 5 4 3 2 1 0 - Printed on acid-free paper

Preface

The adsorption of proteins at interfaces plays a role in many fields, such as health, food, environment and analysis. Fundamental aspects are useful when considering applications. We focus here especially on solid–liquid interfaces and present a few fundamental studies regarding adsorption kinetics and conformational changes, and examples of applications to sensors and membranes. The first part is dedicated to fundamental studies performed using optical waveguide lightmode spectroscopy, as an example of a technique that has the advantage of not requiring labelled proteins, but is limited to specific supports. Conversely, the radiolabelling of proteins, which has the disadvantage of any labelling process, allows application to any kind of surfaces. As proteins bear both positive and negative charges, we can expect the influence of an electric field normal to the interface on the packaging order at interfaces. The refining of data treatment may also lead to the determination of useful structural parameters. The balance between protein–surface and protein–protein interactions is a key point for the description of the structure at high coverage of the surface. Electrokinetic methods, like measurement of the streaming potential, may be helpful in the electrical characterisation of the interfacial layer facing the solution. The second part includes different bench techniques that were developed to improve the sensitivity of the characterisation of the orientation and structure of the proteins at interfaces: dual polarisation interferometry and total internal reflection ellipsometry are such recent examples. Concerning the determination of the secondary structure at interfaces, Fourier transform infrared (FTIR) spectroscopy and circular dichroism are very well adapted. Application to flat model surfaces can be performed by using the attenuated total reflectance-FTIR technique. The accessibility to the internal part of globular proteins is measured by the hydrogen–deuterium exchange rate. The evaluation of proteins on biodevices by time of flight – secondary ion mass spectroscopy is certainly a challenge given the size of the molecules. Data treatment according to mutual information theory, however, might be very helpful. The third part considers studies more closely linked to applications. Determination of the conformation and orientation of the proteins at in-

VI

Preface

terfaces is of particular importance for the understanding of the behaviour of cells at interfaces. A model study is presented with fibronectin at polymer surfaces with graded characteristics such as hydrophilicity, charge density and swelling. In addition to sensors, materials with a large area of contact with solutions containing proteins can be used in large-scale applications. Microporous membranes and textiles are typically representative of this category. They can act as concentrators of enzymes for the highyield transformation of substrates. The biocompatibility of these materials, however, with the specific significance of anti-fouling properties and/or retaining enzyme activity, is a key parameter for satisfactory functioning. Different approaches for obtaining protein-resistant surfaces from polyacrylonitrile membranes are presented. The modulation of the adsorption and activity of biomolecules can also be performed by surface modification of polypropylene membranes. One important “bio-inspired” route to surface modification is the introduction of phosphorylcholine moieties, since the zwitterionic group of the phospatidylcholine and sphingomyelin polar head covers a large fraction of the external surface of the erythrocyte and platelet membranes. I would like to thank the authors for their contributions and the editorial staff of Springer for their assistance. I hope that this book may stimulate initiatives of work in the constantly renewing and fascinating field of interfacial phenomena. Montpellier, February 2006

Philippe Déjardin

Contents

Part I Analysis of the Adsorption Kinetics 1 Protein Adsorption Kinetics: Influence of Substrate Electric Potential Paul R. Van Tassel 1.1 Introduction .................................................................. 1.2 Theoretical Prediction ..................................................... 1.3 Experimental Measure..................................................... 1.3.1 OWLS Principles .................................................. 1.3.2 OWLS Experiments .............................................. 1.4 Results .......................................................................... 1.5 Discussion ..................................................................... 1.5.1 Surface-Bound Counterions ................................... 1.5.2 Local pH Effects ................................................... 1.5.3 Solvent Interfacial Structure .................................. 1.5.4 Protein Charge Heterogeneity ................................ 1.6 Conclusions ................................................................... References ............................................................................. 2 From Kinetics to Structure: High Resolution Molecular Microscopy Jeremy J. Ramsden 2.1 Introduction .................................................................. 2.2 Optical Waveguide Lightmode Spectroscopy ....................... 2.2.1 Principles of Optical Biosensing ............................. 2.2.2 Mode Equations for OWLS .................................... 2.2.3 The Uniform Thin Film Approximation (UTFA) ....... 2.2.4 Optical Invariants................................................. 2.3 The Practical Determination of Waveguide Parameters ........ 2.3.1 Device Fabrication ............................................... 2.3.2 Fluid Handling Arrangements................................ 2.4 Static Structure............................................................... 2.5 Kinetic Analysis and Dynamic Structural Inference ............. 2.5.1 Particle Transport ................................................

1 1 2 6 6 8 9 17 19 20 20 20 21 21 23 23 25 27 28 30 31 34 35 36 37 37 37

VIII

Contents

2.5.2 The Chemical Adsorption Coefficient...................... 2.5.3 The Analysis of The Available Area Function ............ 2.6 Behaviour of Real Proteins ............................................... 2.6.1 Evaluation of Lateral Diffusivity and 2D Crystal Unit Cell Size ................................. 2.6.2 Desorption .......................................................... 2.6.3 Multilayers .......................................................... 2.7 Conclusions ................................................................... References ............................................................................. 3 Initial Adsorption Kinetics in a Rectangular Thin Channel, and Coverage-Dependent Structural Transition Observed by Streaming Potential Philippe Déjardin, Elena N. Vasina 3.1 Introduction .................................................................. 3.2 The Initial Adsorption Constant and its Limit Expressions.... 3.2.1 The Local Initial Adsorption Constant k(x), its Limit Expressions and Approximation................. 3.2.2 The Mean Adsorption Constant, its Limit Expressions and Approximation................. 3.2.3 Experimental Results and Discussion ...................... 3.3 The Structural Transition with Increasing Interfacial Concentration ................................................. 3.3.1 Observation by Streaming Potential ........................ 3.3.2 Different Models .................................................. 3.4 Conclusion .................................................................... Appendix............................................................................... References .............................................................................

40 41 43 44 45 46 47 48

51 51 56 56 59 61 63 64 66 67 68 69

Part II Analysis of the Structure at the Interface 4 Dual Polarisation Interferometry: An Optical Technique to Measure the Orientation and Structure of Proteins at the Solid–Liquid Interface in Real Time Neville Freeman 4.1 Introduction .................................................................. 4.2 Experimental Approaches Adopted ................................... 4.2.1 Typical Approach Adopted .................................... 4.2.2 Experimental Protocols ......................................... 4.2.3 Advantages.......................................................... 4.2.4 Verifying DPI as an Experimental Approach ............

75 75 79 79 79 79 80

Contents

IX

4.3

DPI: Applications............................................................ 80 4.3.1 Introduction........................................................ 80 4.3.2 Protein Orientation .............................................. 81 4.3.3 Bovine Serum Albumin Structures at pH 3 and pH 7 .................................................. 82 4.3.4 Protein Orientation and Subsequent Activity............ 83 4.3.5 Protein Structure and Small Molecule Interactions.... 87 4.3.6 Protein Structure and Metal Ion Interactions............ 90 4.4 Future Developments ...................................................... 91 4.5 Conclusions ................................................................... 93 Appendix 1 DPI: Background ................................................... 93 A.1.1 Neutron Reflection ............................................... 93 A.1.2 Surface Plasmon Resonance ................................... 94 Appendix 2 DPI: Theory .......................................................... 95 Appendix 3 DPI: Implementation .............................................. 99 A.3.1 Hardware ............................................................ 99 A.3.2 Data Analysis....................................................... 101 References ............................................................................. 102 5 Total Internal Reflection Ellipsometry: Monitoring of Proteins on Thin Metal Films 105 Michal Poksinski, Hans Arwin 5.1 Introduction .................................................................. 105 5.2 Total Internal Reflection Ellipsometry................................ 106 5.3 Experimental Setup......................................................... 110 5.4 Application Examples ...................................................... 113 5.5 Further Possibilities ........................................................ 117 References ............................................................................. 118 6 Conformations of Proteins Adsorbed at Liquid–Solid Interfaces 119 Sylvie Noinville, Madeleine Revault 6.1 Introduction .................................................................. 119 6.2 Experimental Techniques ................................................. 125 6.2.1 High-Resolution Structure of Proteins..................... 125 6.2.2 Secondary Structure of Proteins ............................. 126 6.2.3 Orientation, Localised Structural Information .......... 127 6.2.4 Spatial Distribution of Proteins in the Adsorbed Layer ........................................... 128 6.2.5 Solvation Information........................................... 129 6.3 Surface Effects on Both Protein Structure and Solvation by the ATR-FTIR Technique.......................... 130 6.3.1 FTIR Spectral Analysis .......................................... 130

X

Contents

6.3.2 6.3.3

Proteins in Solution .............................................. 132 Surface-Induced Conformational Changes of a Soft Protein: BSA............................................ 134 6.3.4 Surface-Induced Conformational Changes of a Hard Protein: Lysozyme .................................. 138 6.3.5 Folding or unfolding of proteins on hydrophobic supports....................................... 141 6.4 Conclusion .................................................................... 142 References ............................................................................. 142 7 Evaluation of Proteins on Bio-Devices 151 Satoka Aoyagi, Masahiro Kudo 7.1 Introduction .................................................................. 151 7.2 Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS) ................................................................... 153 7.2.1 Principles of TOF-SIMS ......................................... 153 7.2.2 TOF-SIMS Spectra and Secondary-Ion Images .......... 156 7.2.3 Data Analysis....................................................... 157 7.3 Analysis of Proteins on Bio-Devices ................................... 161 7.3.1 Characterization of Proteins on Substrates ............... 161 7.3.2 Investigation of Conformation and Orientation of Proteins on Substrates ....................................... 164 7.3.3 Imaging of Protein Distribution ............................. 165 7.3.4 Other Points and Future Directions......................... 168 7.4 Summary....................................................................... 169 References ............................................................................. 169

Part III Some Applications 8 Fibronectin at Polymer Surfaces with Graduated Characteristics 175 Tilo Pompe, Lars Renner, Carsten Werner 8.1 Introduction .................................................................. 175 8.2 Gradated Substrate Physicochemistry ................................ 177 8.3 Fibronectin Exchange at a Constant Surface Concentration ... 181 8.4 Fibronectin Exchange at Variable Surface Concentrations ..... 188 8.5 Relevance of the Interfacial Constraints of Fibronectin for Cell-Matrix Adhesion............................. 195 References ............................................................................. 197

Contents

XI

9 Development of Chemical Microreactors by Enzyme Immobilization onto Textiles 199 Christophe Innocent, Patrick Seta 9.1 Introduction .................................................................. 199 9.2 Nonconducting Cellulosic Textiles ..................................... 201 9.2.1 Pepsin and Trypsin Immobilization on Cotton.......... 201 9.2.2 Immobilization of Uricase and Xanthine Oxidase on Ion-Exchanging Textiles.................................... 211 9.2.3 Urease Electrodialysis Coupling ............................. 223 9.3 Electron-Conducting Textile............................................. 227 9.3.1 Enzyme Immobilization on Carbon Felt .................. 227 9.3.2 Electrocatalysis Coupling with Enzyme-Conducting Textile Catalytic Reactivity .................................... 238 References ............................................................................. 242 10 Approaches to Protein Resistance on the Polyacrylonitrile-based Membrane Surface: an Overview 245 Ling-Shu Wan, Zhi-Kang Xu, Xiao-Jun Huang 10.1 Introduction .................................................................. 245 10.2 Copolymerization Procedures........................................... 246 10.3 Poly(ethylene glycol) Tethering ......................................... 252 10.4 Physical Adsorption ........................................................ 257 10.5 Biomacromolecule Immobilization.................................... 259 10.6 Biomimetic Modification ................................................. 263 10.7 Conclusion .................................................................... 266 References ............................................................................. 268 11 Modulation of the Adsorption and Activity of Protein/Enzyme on the Polypropylene Microporous Membrane Surface by Surface Modification 271 Qian Yang, Zhi-Kang Xu, Zheng-Wei Dai 11.1 Surface Modifications for Reducing Nonspecific Protein Adsorption ......................................................... 271 11.1.1 Plasma treatment ................................................. 273 11.1.2 Ultraviolet (UV) modification ................................ 276 11.1.3 γ -Ray-induced modification .................................. 282 11.1.4 Ozone Method ..................................................... 285 11.2 Surface-Modified PPMMs for Enzyme Immobilization ......... 286 11.2.1 Physical Adsorption/Entrapment............................ 287 11.2.2 Covalent Binding.................................................. 289 11.2.3 Site-Specific Immobilization .................................. 294 11.3 Conclusions ................................................................... 295 References ............................................................................. 295

XII

Contents

12 Nonbiofouling Surfaces Generated from Phosphorylcholine-Bearing Polymers 299 Yasuhiko Iwasaki, Nobuo Nakabayashi, Kazuhiko Ishihara 12.1 Introduction .................................................................. 299 12.2 Forces Involved in Protein Adsorption ............................... 300 12.3 Design of Phosphorylcholine-Bearing Surfaces ................... 302 12.4 Mechanism of Resistance to Protein Adsorption on the MPC Polymer Surface ............................................ 303 12.5 Fundamental Interactions Between MPC Polymers and Proteins................................................................... 310 12.6 Recent Designs of Nonfouling Phosphorylcholine Surfaces with Well-Defined Structures ............................... 312 12.7 Control of Cell–Material Interactions on a Phosphorylcholine Polymer Nonfouling Surface ........... 314 12.7.1 Cell Manipulation on a Well-Defined Phosphorylcholine Polymer Brush.......................... 315 12.7.2 Selective Cell Attachment to a Biomimetic Polymer Surface Through the Recognition of Cell-Surface Tags .......... 318 12.8 Conclusion .................................................................... 321 References ............................................................................. 321 Subject Index

327

Contributors

Aoyagi, Satoka Faculty of Life and Environmental Science, Department of Regional Development, Shimane University 1060 Nishikawatsu-cho, Matsue-shi, Shimane, 690-8504, Japan, Shimane University, Japan Arwin, Hans Laboratory of Applied Optics, Department of Physics, Chemistry and Biology, Linköping University, SE-581-83 Linköping, Sweden Déjardin, Philippe European Membrane Institute, UMR 5635 (ENSCM-UMII-CNRS), Université Montpellier 2, CC 047, 34095 Montpellier Cedex 5, France Dai, Zheng-Wei Institute of Polymer Science, Zhejiang University, Hangzhou 310027, People’s Republic of China Freeman, Neville Farfield Scientific Ltd, Farfield House, Southmere Court, Electra Way, Crewe Business Park, Crewe, Cheshire, CW1 6GU, UK Huang, Xiao-Jun Institute of Polymer Science, Zhejiang University, Hangzhou 310027, People’s Republic of China Innocent, Christophe European Membrane Institute, UMR 5635 (ENSCM-UMII-CNRS), Université Montpellier 2, CC 047, 34095 Montpellier Cedex 5, France Ishihara, Kazuhiko Department of Materials Engineering, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan

XIV

Contributors

Iwasaki, Yasuhiko Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, 2-3-10 Kanda-surugadai, Chiyoda-ku, Tokyo 101-0062, Japan Kudo, Masahiro Department of Materials and Life Science, Seikei University, Japan Nakabayashi, Nobuo Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, 2-3-10 Kanda-surugadai, Chiyoda-ku, Tokyo 101-0062, Japan Noinville, Sylvie Laboratoire de Dynamique, Interactions et Réactivité, CNRS-Université Pierre et Marie Curie, UMR 7075, 2 rue Henri Dunant, 94320 Thiais, France Poksinski, Michal Laboratory of Applied Optics, Department of Physics, Chemistry and Biology, Linköping University, SE-581-83 Linköping, Sweden Pompe, Tilo Leibniz Institute of Polymer Research Dresden, Max Bergmann Center of Biomaterials Dresden, Hohe Str. 6, 01069 Dresden, Germany Ramsden, Jeremy J. Department of Materials, Cranfield University, MK43 0AL, UK Renner, Lars Leibniz Institute of Polymer Research Dresden, Max Bergmann Center of Biomaterials Dresden, Hohe Str. 6, 01069 Dresden, Germany Revault, Madeleine Laboratoire de Dynamique, Interactions et Réactivité, CNRS-Université Pierre et Marie Curie, UMR 7075, 2 rue Henri Dunant, 94320 Thiais, France Seta, Patrick European Membrane Institute, UMR 5635 (ENSCM-UMII-CNRS), Université Montpellier 2, CC 047, 34095 Montpellier Cedex 5, France Van Tassel, Paul R. Department of Chemical Engineering, Yale University, New Haven, CT 06520-8286, USA

Contributors

XV

Vasina, Elena N. European Membrane Institute, UMR 5635 (ENSCM-UMII-CNRS), Université Montpellier 2, CC 047, 34095 Montpellier Cedex 5, France Wan, Ling-Shu Institute of Polymer Science, Zhejiang University, Hangzhou 310027, People’s Republic of China Werner, Carsten Leibniz Institute of Polymer Research Dresden, Max Bergmann Center of Biomaterials Dresden, Hohe Str. 6, 01069 Dresden, Germany Xu, Zhi-Kang Institute of Polymer Science, Zhejiang University, Hangzhou 310027, People’s Republic of China Yang, Qian Institute of Polymer Science, Zhejiang University, Hangzhou 310027, People’s Republic of China

1

Protein Adsorption Kinetics: Influence of Substrate Electric Potential Paul R. Van Tassel

Abstract. Substrate electric potential plays an important role in determining the adsorptive behavior of proteins and other macromolecules. In this chapter we describe the measurement of protein adsorption kinetics in the presence of an applied potential using optical waveguide lightmode spectroscopy. We analyze the resulting kinetics in terms of transport- and surface-limited models and show that while substrate potential is an important influencing factor, transport limited by convective diffusion and adsorption in seeming violation of electrostatic principles are prevalent in simple protein systems.

1.1 Introduction Protein molecules immobilized at a material surface play a key role in many biosensing, tissue engineering, enzymatic catalysis, bioseparation, and bioelectronics applications. The tendency of proteins to attach to interfacial regions is well documented (Haynes and Norde 1994; Malmsten 1998a, b; Van Tassel 2003). Ionic, van der Waals, solvation, and donor– acceptor interactions all play important roles in rendering the interfacially adsorbed state to be thermodynamically favored over the solution state (Haynes and Norde 1994). Proteins are colloidal objects that possess a distribution of surface charge and, in an electrolytic solution, a distribution of weakly associated counterions. Their interaction with a solid substrate is thus sensitive to the substrate’s charge distribution. By controlling the electric potential of an adsorbing surface, one alters this charge distribution and, therefore, the surface–protein interaction. In principle, the adsorption process may be controlled in this way, perhaps leading to adsorbed layers of enhanced density or preferred orientation or spatial distribution. However, the interaction between a protein and an adsorbing surface is complex, and predicting adsorbed layer properties by considering the contributions from Paul R. Van Tassel: Department of Chemical Engineering, Yale University, New Haven, CT 06520-8286, USA, E-mail: [email protected] Principles and Practice Proteins at Solid–Liquid Interfaces Philippe Déjardin (Ed.) © Springer-Verlag Berlin Heidelberg 2006

2

P.R. Van Tassel

the interaction modes given above remains a significant challenge. Thus, while influencing an adsorbed protein layer through the substrate’s electric potential is both possible and desirable (Asanov et al. 1997, 1998; Bernabeu and Caprani 1990; Bos et al. 1994; Brusatori et al. 2003; Brusatori and Van Tassel 2003; Feng and Andrade 1994; Fievet et al. 1998; Fraaije et al. 1990; Khan and Wernet 1997), the process is as yet poorly understood. In this contribution, we review our recent work investigating the influence of substrate electric potential on protein adsorption kinetics (Brusatori et al. 2003; Brusatori and Van Tassel 2003). We begin with a brief presentation of certain basic theoretical considerations. We then introduce the method by which we measure protein adsorption kinetics under electric potential control: optical waveguide lightmode spectroscopy (OWLS). Next, we introduce some key results and discuss them in the context of findings by other groups. By summarizing several important results and open questions, we hope to guide future efforts to produce controlled layers of adsorbed protein through the control of substrate electric potential.

1.2 Theoretical Prediction Proteins are composed of amino acids, some of which contain acidic/basic sites. Thus, at all but the isoelectric pH, a protein molecule possesses a net charge and therefore migrates in response to an electric field. Since the charge distribution is generally not spherically symmetrical, the electric field also imposes a torque on the molecule, causing it to rotate. Of course, electric-field-induced migration and rotation must compete against the molecule’s thermal diffusive motion, so these influences are only observed in excess of some threshold field strength. Due to screening by solution ions, electric field effects become appreciable only within a few Debye lengths of a charged surface and are most important at or near direct contact, where properties such as the rate of attachment, mean protein orientation, and the rate and extent of subsequent postadsorption changes in orientation and conformation may be profoundly affected. The transport of protein in a flowing solution is described by → → cq E ∂c → → → + v • ∇c = ∇ • D∇c + (1) ∂t ζ →

where c is the protein concentration, t is the time, v is the fluid → flow velocity vector, D is the diffusivity, q is the effective protein charge, E is the electric field vector, and ζ is the friction coefficient (a measure of the viscous drag

1 Protein Adsorption Kinetics: Influence of Substrate Electric Potential

3

Fig. 1. A schematic of a typical protein adsorption scenario, where adsorption occurs from a flowing solution over a flat, charged surface. Near the surface, simple shear flow occurs → resulting in a velocity profile v = azˆx, where a is the shear rate, z is the distance from the surface, and xˆ is a unit vector parallel to the surface. A convective-diffusion boundary convection→is more (less) rapid layer develops δ(x, t), of thickness δ, above (below) which → than diffusion. An applied electric field is expressed as E = Eˆz, where E is the electric field vector, E is the magnitude and zˆ is a unit vector normal to the surface

on the protein). The flow field must be known in advance to solve Eq. 1. So long as the concentration is quite low, one may neglect the influence of the migrating proteins on the fluid and solve for the flow field using the standard Navier-Stokes equation. A typical configuration involves flow past a flat, charged surface, as depicted in Fig. 1. The two terms on the right side of Eq. 1 represent, respectively, the contributions from thermal diffusion and electric-field-induced migration. It is interesting to consider the limiting cases where one is much larger than the other. Neglecting the second term gives pure convective diffusion. In this case, one may approximate the flux (J) to the surface by assuming steady shear flow over a flat surface acting as a perfect adsorptive sink, J(x, t) = Dcb /δ(x, t), where cb is the bulk protein concentration, and (45Dx/ 2a)1/ 3 4Dt f (45Dx/ 2a)2/ 3

δ(x, t) =

(2)

is the convective diffusion boundary layer, a function of the diffusivity (D), the distance along the surface in the direction of the flow (x), the shear rate (a), and the time (t) (Brusatori et al. 2003; Calonder and Van Tassel 2001). The function f (τ) is the inverse of τ(f ) = [1 − (1 − f 3 )2/ 3 ]/ 2f 2 for τ < 1/ 2 and f (τ) = 1 for τ > 1/ 2. Pure convective diffusion is therefore characterized by an initially zero flux that increases to a steady-state value at dimensionless time, τ = 1/ 2. Inclusion of the second term on the right of Eq. 1 requires knowledge of the electric field as a function of position. A starting point toward its approximation is the Poisson-Boltzmann equation describing the distribution of electric potential in an ionic solution: e 0 −zi eΨ/ kT ∇ 2Ψ = − zi ci e (3)

ε

i

4

P.R. Van Tassel

In Eq. 3, Ψ is the electric potential, e is the elementary charge, ε is the dielectric constant, zi is the valance of the ith ionic species, ci0 is the bulk concentration of the ith species, k is the Boltzmann constant, and T is the absolute temperature. An important assumption leading to Eq. 3 is the complete neglect of spatial correlations among the ions in solution. Clearly, the validity of this assumption is suspect for higher ionic concentrations. When the electric potential energy is weaker than the thermal energy, the exponential in Eq. 3 may be linearized. The result is the linear PoissonBoltzmann equation, ∇ 2 Ψ = κ2 Ψ , where κ is the inverse of the Debye length, ⎡ 2 2 0 ⎤1/ 2 e zi ci i ⎦ ⎣ κ= (4) εkT The linearized version of Eq. 3 may be solved over the region extending away from a flat surface to yield

Ψ (x) = Ψ0 e−κz =

σ −κz e εκ

(5)

where Ψ 0 is the surface potential, z is the distance from the surface, and σ is the surface charge density (the second equality follows from integrating the charge density as given by the Poisson equation). The electric field is then given by → dΨ σ zˆ = e−κz zˆ E(x) = − (6) dz ε where xˆ is the unit vector normal to the flat surface. Neglecting the first term on the right side of Eq. 1 is equivalent to ignoring diffusive motion. Assuming fully developed shear flow, a no-slip, perfect sink boundary condition, and an electric field given by Eq. 6, the z component velocity and the z position of a protein are given by: qE(z) qσ −κz dz qσκt −κz0 −1 vz = ⇒ z(t) = z0 + κ ln 1 + = e = e ζ εζ dt εζ (7) The velocity in the x direction is vx (t) = az(t), so the distance traveled in the x direction (x = 0 corresponds to the leading edge of the surface, as shown in Fig. 1) is: t

x(t) = vx t dt (8) 0 εζ σκ t q = az0 t + aκ−1 t + eκz0 ln 1 + e−κz0 − t qσκ εζ


5

The concentration profile is thus given by: qσκt −κz0 εζ κz0 −1 e e t+ ln 1 + −t −x c(x, z, t) = cb H az0 t + aκ qσκ εζ (9) where z0 is the implicit solution to Eq. 7, in terms of z(t) and t, and H is the Heaviside function (i. e., H(λ) = 1 for λ > 0 and H(λ) = 0 for λ < 0). The flux to the surface is thus: qσ c(x, z = 0, t) (10) J(x, t) =

εζ

Pure electrophoretic migration is therefore characterized by an initial period of zero flux, during a time needed for the argument of the Heaviside function in Eq. 9 to vanish, followed by a steady-state flux. Both of these flux predictions hold only in the limit where surface adsorption is rapid compared to transport to the surface (i. e., where the surface may be considered to be a perfect sink). This is usually the case during the initial stages of surface filling. Subsequently, the kinetics are limited by surface effects. The surface-limited rate of adsorption may be expressed as: dΓ = ka cb Φ − k d,i Γi (11) dt i where Γ is the density of adsorbed protein (mass per area), ka is the adsorption rate constant, Φ is the one-body cavity function, and kd,i and Γi are the desorption rate constant and the density of protein in the ith structural state, respectively (these states may denote various conformations, orientations, or states of aggregation; Tie et al. 2003). The cavity function is defined as Φ = e−u/ kT Γ,T , where u is the potential energy of a single molecule interacting with the surface and with all of the previously adsorbed molecules (u depends on position and orientation), k is the Boltzmann constant, T is the absolute temperature, and the brackets represent an averaged quantity, over all representations of the adsorbed layer at density Γ and temperature T, according to their appropriate weights, and over all orientations and positions of the single “reference” molecule. All of the quantities on the right of Eq. 11 (except cb ) may be altered by application of an electric field. To make quantitative predictions, the potential energy of interactions between protein molecules and the charged surface must be calculated. The electrostatic contribution to this energy may be determined using the Derjaguin-Landau-Verwey-Overbeek (DLVO; Asthagiri and Lenhoff 1997; Oberholzer and Lenhoff 1999; Ravichandran and Talbot 2000; Roth and Lenhoff 1993, 1995) or density functional (Carignano and Szleifer 2002; Fang and Szleifer 2003) approaches.

6

P.R. Van Tassel

1.3 Experimental Measure 1.3.1 OWLS Principles OWLS is a method for detecting interfacial adsorption based on the guided light modes excited in a planar, dielectric waveguide sandwiched between media of lower refractive index (Ramsden 1993; Voros et al. 2002). The OWLS detection mechanism may be understood through a four-layer model, consisting of a glass support (S), a waveguiding film (F), a second film (F ), and a cover solution (C)(Brusatori and Van Tassel 2003; Tiefenthaler and Lukosz 1989). In what follows, the F layer may represent an adsorbed layer or a second layer to a two-layer waveguiding film. A polarized laser beam of wavelength λ is directed through the glass support toward an optical grating at the FF C interface at an angle α. The corresponding effective refractive index (N) is: N ≡ nF sin φF = nair sin α +

λ

Λ

(12)

where nF is the film’s refractive index, φF is the propagation angle within the film, nair is the refractive index of air, is the diffraction order of the grating, and Λ is the grating period. The condition under which coherent propagation occurs in the direction parallel to the film is: 2kz,F dF + ϕF,S + ϕF,F ,C = 2πm

(13)

where kz,F is the component of the wave vector normal to the waveguiding film, dF is the thickness of the waveguiding film, ϕF,S and ϕF,F , C are the phase shifts associated with the reflections at the F,S and the F,F ,C interfaces, respectively, and the integer m is the mode number. The phase shift at a given interface is related to its reflection coefficient via r = |r|eiϕ . For a simple interface between films α and β, kz,α 2ρ

rα,β =

nα

kz,α 2ρ

nα

− +

kz,β 2ρ

nβ kz,β

(14)

2ρ

nβ

where α, β = F, F , or C, nα is the refractive index of film α, and kz,α is the component of the wave vector normal to the film α: kz,α =

2π

λ

nα cos φα

(15)


7

where φα is the propagation angle in the α film (by Snell’s law, cos φα = [1−N 2 /Re(nα )2 ]1/ 2 is the real part of nα . In Eq. 14, ρ = 0 or 1 for electric field vectors oriented normal (transverse electric, or TE) or parallel (transverse magnetic, or TM) to the plane of incidence, respectively. The reflection coefficient across the F,F ,C interface is: rF,F ,C =

rF,F + rF ,C eiδ 1 + rF,F rF ,C eiδ

(16)

where

δ = 2kz,F dF +

4πi

λ

Im(nF )dF tan φF sin φF

(17)

and Im(nF ) is the imaginary component of the refractive index. The phase shifts ϕF,S and ϕF,F ,C are expressed in terms of the associated reflection coefficients by ϕ = arcsin[Im(r)/ |r|]. In the absence of an F layer, the above approach allows for the solution of the refractive index and thickness of the waveguiding film, nF and dF from the effective refractive indices of the transverse electric (NTE ) and transverse magnetic (NTM ) modes and from the known refractive indices of the support (nS ) and cover (nC ) solutions. Once known, these quantities may be used to calculate nF and dF during an adsorption experiment. From these values, the mass density of an adsorbed layer is determined via de Feijter’s formula (de Feijter et al. 1978):

nF − nC dF Γ= (18) dn C

dcb

where the refractive index increase with bulk protein concentration, dnC / dcb , has the nearly universal value of 0.182 cm3 /g. Alternatively, the mass density may be determined via a single effective refractive index by:

∂nC −1 Γ = nF − nC dF ∂cb (19) 2 2 nF − nC ∂NTE −1 ∂nC −1 = ∆NTE nF + nC ∂dF ∂cb where the refractive index nF and the term ∂NTE /∂dF are determined by a separate double mode experiment. The angular scan required for a singlemode measurement is much narrower, so the rate at which measurements may be taken (approximately 3 s) greatly exceeds that of a double-mode experiment (approximately 20 s).

8

P.R. Van Tassel

1.3.2 OWLS Experiments A schematic of our OWLS system (BIOS-1, MicroVacuum, Budapest, Hungary), as modified to allow for detection in the presence of an applied potential difference, appears in Fig. 2. The adsorbing surface is an indium tin oxide (ITO, or In2−2x Snx O3−x with x = 0.50 ± 0.02)-coated optical waveguide 2400 sensor chip (MicroVacuum). The ITO layer has a thickness of approximately 10 nm and complex refractive index 1.80−0.03i and rests on a silicon titanium oxide layer (STO, or Si1−x Tix O2 with x = 0.25 ± 0, 05) of thickness ca. 200 nm and refractive index 1.77 ± 0.03. The water contact angle of the ITO coating is 48.1 ± 0.1◦ ; this value does not change within the range of applied potential investigated here. The sensor chip serves as the base of a temperature-controlled flow cell of volume 70 µl. A flat platinum counterelectrode is situated at the top of the flow cell 1.0 mm from the ITO surface. The sensor chip/flow cell assembly rests on the head of a precision goniometer. An electric circuit schematic appears in Fig. 3. A potential difference between the ITO working electrode and the platinum counter is applied via an external power supply. The current is measured across a 100-kW resistor via a voltmeter, and the magnitude of the potential difference is determined using a second voltmeter. The potential of the ITO surface relative to a reference electrode (situated in the inlet solution) is measured using an electrometer.

Fig. 2. A schematic of our modified optical waveguide lightmode spectroscopy (OWLS) system for the continuous measurement of macromolecular adsorption under an applied potential. The adsorbing substrate is an approximately 10-nm indium tin oxide (ITO) layer on a silicon titanium oxide (STO) waveguiding film of approximately 200 nm, itself supported on a glass substrate (not shown). A potential difference (∆V) is applied between the ITO layer and a flat platinum (Pt) counterelectrode, situated 1 mm above at the top of the flow cell. A polarized HeNe laser beam directed toward a grating coupler at the STO/ITO interface (not shown) from below excites a guided mode at a resonant angle, from which adsorbed layer mass and thickness may be determined


9

Fig. 3. A schematic of the electric circuit of our OWLS system. The ITO– and Pt–solution ITO Pt interfaces may be considered as resistors (RITO int and Rint , respectively) and capacitors (Cint Pt and Cint , respectively) in parallel. The ITO film and the solution within the flow cell act as resistors (RITO and Rsol , respectively). A potential difference is applied through a power supply (V) and the current is measured via the voltage drop across a 100-kΩ resistor. A second voltmeter is used to measure the potential difference between the ITO and Pt electrodes. The potentials of the ITO and Pt electrodes versus a reference electrode (e. g., Ag/AgCl) are measured using an electrometer (EM)

Prior to each experiment, the flow cell, tubing, and sensor chip are cleaned by exposure to a 2% Hellmanex (Hellma, Mulheim, Germany) solution in ultrapure water, followed by an intensive rinse with ultrapure water. An experiment begins with the introduction of a pure buffer to the flow cell via a peristaltic pump. An angular scan is performed, resulting in a mode spectrum from which NTE and NTM are obtained at approximately 20-s intervals. During an open-circuit potential (OCP) experiment, a flowing protein solution then replaces the flowing buffer solution once stable values of NTE and NTM are achieved. In other experiments, a voltage difference ∆V is applied (∆V = VITO − VPt ), resulting in further changes to NTE and NTM . Once steady values are recovered, a protein solution is introduced.

1.4 Results We begin by discussing certain important optoelectronic properties of the ITO-coated sensor chip. In Fig. 4, we present the effective refractive index, the current, and the ITO and platinum electrode potentials versus applied voltage in the presence of deionized water (of pH 5.5–6 and conductivity 1.30 ± 0.05 µS at 25 ◦ C) and HEPES buffer [10 mM N-(2-hydroxyethyl)

10

P.R. Van Tassel

Fig. 4. The asymptotic current, the electric potential of the ITO and Pt electrodes, and the effective refractive index of the ITO-coated sensor chip as a function of applied voltage for a water and b HEPES buffer [10 mM N-(2-hydroxyethyl) piperazine-N -(ethanesulfonic acid) in 100 mM NaCl, adjusted to pH 7.4 by addition of 6 N NaOH] solvents. The increases in current at 1.5 V and 1.0 V, respectively, indicate the onset of water electrolysis. NTE Effective refractive index of the transverse electric mode. Reproduced with permission from Brusatori et al. (2003)

piperazine-N -(ethanesulfonic acid) in 100 mM NaCl, adjusted to pH 7.4 by addition of 6 N NaOH]. We observe the effective refractive index to increase with applied potential up to ∆V = VITO − VPt > 1.4 V. Upon further voltage increase, NTE decreases in the water system and plateaus in the HEPES system. These changes suggest an alteration in optical density of the ITO-coated waveguiding film due to accumulation of charged species at the interface and/or within the film, and possibly also to mild oxidative reactions. A voltage-induced orientation of water in the porous ITO


11

film is also a possible explanation, as suggested elsewhere (Stankowski and Ramsden 2002). The somewhat sharp increase in current above an applied voltage of 1.5 V (water) and 1.0 V (HEPES) is due to the onset of water electrolysis. However, we observe no H2 or O2 gas formation and attribute this absence to the low overall current density.∗ Increasing the applied voltage increases VITO and decreases VPt ; the latter becomes negative relative to a standard hydrogen electrode (SHE) when ∆V exceeds 1.7 V and 1.5 V for water and HEPES, respectively. We find prolonged cathodic polarization (i. e., ∆V < 0) to result in increased opacity of the ITO film. This is probably due to electrochemical reduction and renders the sensor chip unusable in an OWLS experiment. We wish to use the measured effective refractive index to calculate the adsorbed protein density, but doing so brings forth two important questions: (1) Must the ITO layer be considered explicitly, and if so, must the imaginary portion of its complex refractive index be considered? (2) Do the model Eqs. 12–19 hold under an applied voltage? To answer these questions, we measure the refractive index of a (non-adsorbing) 5.0 mg/ml glucose solution in water, with and without an applied potential, using three different waveguide models: (1) a three-layer model in which the STO and ITO films are considered to be a single layer (i. e., both reflection coefficients are determined via Eq. 12), (2) a four-layer model in which nF is real (i. e., the F layer is a dielectric), and (3) a four-layer model in which nF is complex (i. e., the F layer is a conductor). The glucose solution has a known refractive index of 1.33173 ± 1×10−5 at 25 ◦ C. In Table 1, we show refractive index values measured using OWLS within each of the three waveguide models for cases of OCP and an applied voltage of ∆V = 5 V. We observe all values to be within 3×10−5 of one another. Since the difference (nF − nC ) appearing in Eq. 18 is typically around 0.5, uncertainty of this magnitude has no appreciable effect on a measured adsorbed density. Thus, we conclude that an applied potential difference has only a negligible effect on the instrument’s detection and that one may consider the STO and ITO films as a single effective film. Using OWLS and employing a four-layer waveguide model – in which the STO and ITO layers are treated as a single layer, as justified above – we measure the adsorption kinetics of human serum albumin and horse heart cytochrome c, in flowing water or HEPES solutions at a surface shear rate of 1.5/s, onto ITO as a function of applied potential (Brusatori et al. 2003; Brusatori and Van Tassel 2003). The raw data of a typical experiment are shown in Fig. 5. Following a baseline measured under flowing solvent (deionized water in this case), application of a potential difference (1.0 V) ∗ A current of 1 µA results in the production of approx. 0.5×10−7 M/s of H in our flow 2 cell. The time needed to achieve the solubility limit of approx. 10−3 M is thus approximately 2×104 s; this is much greater than the space time of the flow cell, which is about 50 s.

12

P.R. Van Tassel

Table 1. The refractive index, nC , of a 5.0 mg/ml glucose solution – for an open circuit potential (OCP) and an applied voltage of 5 V – as determined using (1) a three-layer waveguide model where the silicon titanium oxide (STO) and indium tin oxide (ITO) layers are treated as a single layer, (2) a four-layer waveguide model where the ITO film is treated as a separate dielectric layer (i. e., of real refractive index), and (3) a four-layer waveguide model where the ITO film is treated as a separate conducting layer (i. e., of complex refractive index). The closeness of the values demonstrates both the validity of treating the STO and ITO films as a single optical layer and the accuracy of detection in the presence of an applied voltage OCP OCP

Three-layer waveguide model with composite nC = 1.33177 ± 1×10−5 STO-ITO layer Four-layer waveguide model with dielectric ITO layer nC = 1.33178 ± 1×10−5

Four-layer waveguide model with conducting ITO nC = 1.33177 ± 1×10−5 layer ∆V = 5 V Three-layer waveguide model with composite nC = 1.33175 ± 2×10−5 STO-ITO layer ∆V = 5 V Four-layer waveguide model with dielectric ITO layer nC = 1.33175 ± 2×10−5 OCP

∆V = 5 V Four-layer waveguide model with conducting ITO

nC = 1.33175 ± 2×10−5

layer

Fig. 5. The effective refractive index (the raw output signal during an OWLS measurement) and the current during cytochrome c adsorption from water onto ITO as a function of time. At point i, a voltage difference of 1.0 V is applied. At point ii, the protein solution is introduced. At point iii, the protein solution is replaced by a buffer solution. At point iv, the system is returned to an open-circuit potential (OCP), and at point v, the 1.0 V voltage difference is reapplied. At point vi, the protein solution is reintroduced. Reproduced with permission from Brusatori et al. (2003)

between the adsorbing substrate and a platinum counterelectrode (i) yields an increase in the measured effective refractive index, the fundamental output signal of OWLS (see Eq. 10). This increase is not fully understood, but is likely to be due to penetration of small ions into the ITO and/or the


13

underlying silicon titanium oxide films (even in deionized water, carbonate and other ionic impurities are present), or to mild oxidation of the ITO. The current increases initially but then decreases, as would occur for a resistor and capacitor in series (see Fig. 3). Upon the addition of cytochrome c in water (ii), the signal greatly increases due to optical changes at the interface brought about by adsorption. A return to the deionized water (iii) causes only a small signal reduction, indicating that only a small quantity of protein desorbs. When the electric field is removed (iv), a decreased signal results; however, this is not due to further desorption as a re-application of the field (v) returns the signal to its previous level. Finally, a re-introduction of the protein (vi) yields an additional signal increase. This additional adsorption is likely to be due to an increased amount of area on the surface open for adsorption. The cause of this increased available area is probably aggregation among the adsorbed molecules. In Figs. 6 and 7, we show the adsorbed density versus time for albumin and cytochrome c in both water and HEPES solvents. In all cases, we observe the total amount adsorbed to increase with increasing applied potential. For water systems, adsorption curves reach a virtual plateau during OCP (∆VOCP = −0.02 ± 0.01 V) but fail to do so in the presence of an applied voltage. The amount adsorbed during 1 h is greatly increased (by up to a factor of 4) by an applied potential difference. We observe a similar, but less pronounced voltage dependence in the HEPES systems (where ∆VOCP = +0.02 ± 0.01 V). Variability is approximately 20% for runs using different ITO-coated sensor chips (most likely due to subtle variations in surface microstructure) and approximately 8% for runs on the same sensor chip. The principal advantage of OWLS over other detection methods is the precision of the output signal. We exploit this advantage in Fig. 8 by plotting the discrete time derivative of the adsorbed density, dΓ/dt, as a function of surface density, Γ. These derivative data are themselves sufficiently precise so as to allow for a clear delineation of transport- and adsorption-limited regimes. Adsorption is initially limited by the rate of transport to the surface; this occurs by a combination of diffusion and electrophoretic migration in the presence of a flow field. We observe the adsorption rate to continuously increase during this regime, as would occur initially for convective diffusion in laminar flow (see Theoretical Prediction). If the surface were a perfect sink, a plateau would eventually be reached upon achievement of a steady-state concentration profile above the surface. In the example shown here, surface availability begins to reduce the rate of adsorption before such a steady, transport-limited rate is achieved. Theoretical treatments of the surface-limited regime (Eq. 11) predict an initial linear decrease in adsorption rate with increasing surface coverage, followed by a nonlinear asymptotic dependence; our observations are consistent with

14

P.R. Van Tassel

Fig. 6. The density of adsorbed protein versus time for a albumin and b cytochrome c in water onto ITO at OCP and under various applied potentials (1.0, 1.5, and 2.0 V). Since monolayer densities are approximately 0.46 and 0.33 µg/cm2 for albumin and cytochrome c, respectively, multilayer adsorption is occurring under an applied potential

these predictions. Extrapolation of data in the linear portion of the surfacelimited regime to zero density yields an estimate of the apparent adsorption rate constant, ka (Calonder et al. 2001; Calonder and Van Tassel 2001). Rate curves such as these also allow for an estimate of the free protein concentration at the surface, and thus of the perfect sink approximation employed in the analysis shown in “Theoretical Prediction”. For a given adsorbed density, this is just the ratio of the adsorption rate to the corresponding value of the extrapolated linear region.


15

Fig. 7. The density of adsorbed protein versus time for a albumin and b cytochrome c in HEPES buffer onto ITO at OCP and under various applied potentials (1.0 and 1.5 V)

In Fig. 8, we show the adsorption rate as a function of the adsorbed amount (and, in the insets, as a function of time) for albumin and cytochrome c in water. In the insets, we also show predictions based upon purely convective diffusion (Eq. 2). For albumin, we observe the adsorption rate to increase with applied voltage over the entire range of surface coverage. As shown by the inset, the initial rate is increased significantly by the presence of an applied potential, indicating a transport mechanism to which both convective diffusion and electrophoretic migration contribute. It should be noted that the observed behavior is far from the limit of

16

P.R. Van Tassel

Fig. 8. The rate of adsorption versus adsorbed density (and in the inset, versus time) for a albumin and b cytochrome c in water onto ITO at OCP and under various applied voltages (1.0, 1.5, and 2.0 V). Also shown are predictions (solid lines) for pure convective diffusion given by Eq. 2. Both convective diffusion and electromigration influence transport of the negatively charged albumin, but only convective diffusion influences the transport of the positively charged cytochrome c. If electromigration were the dominant transport mechanism, an initial nonzero adsorption rate would result, as predicted by Eq. 10. Γ Surface density, dΓ/ dt time derivative of the adsorbed density, t time

pure electrophoretic migration characterized by an initially nonzero, timeindependent adsorption rate, as predicted by Eq. 10. For cytochrome c, we observe the transport-limited rates to be essentially independent of ap-


17

Table 2. The apparent adsorption rate constant ka determined by extrapolating the linear region of a time derivative of the adsorbed density ( dΓ/ dt) versus surface density (Γ) curve (see Fig. 8), and the first-order density expansion coefficient to the cavity function C1 (Eq. 11), for human serum albumin and horse heart cytochrome c (Cyt c) in water and HEPES [10 mM N-(2-hydroxyethyl) piperazine-N -(ethanesulfonic acid) in 100 mM NaCl, adjusted to pH 7.4 by addition of 6 N NaOH] buffer onto ITO at an OCP and under an applied voltage (1.0, 1.5, or 2.0 V) Protein

Applied potential (V)

H2 O solvent C1 ka (10−5 cm/s) (cm2 /µg)

HEPES solvent ka C1 (10−5 cm/s) (cm2 /µg)

Albumin

OCP 1.0 1.5 2.0

1.5 ± 0.1 3.4 ± 0.1 5.2 ± 0.2 9.1 ± 0.2

−3.6 ± 0.1 −2.8 ± 0.1 −3.2 ± 0.3 −2.1 ± 0.1

0.4 ± 0.01 1.1 ± 0.1 1.5 ± 0.1 –

−9.2 ± 0.6 −8.2 ± 1.0 −7.1 ± 0.7 –

Cyt c

OCP 1.0 1.5 2.0

9.9 ± 0.3 9.5 ± 0.2 9.0 ± 0.3 8.3 ± 0.3

−4.5 ± 0.4 −4.7 ± 0.1 −4.5 ± 0.3 −4.3 ± 0.4

3.8 ± 0.2 4.5 ± 0.3 4.3 ± 0.6 –

−8.1 ± 0.7 −6.7 ± 0.8 −6.7 ± 2.0 –

plied potential and the surface-limited rates to be enhanced only at higher coverage (i. e., in the asymptotic region). In Table 2, we show the apparent adsorption rate constant, ka , versus voltage obtained by extrapolating data in the linear regions. For albumin in both water and HEPES solvents, we find ka to increase significantly with voltage. For cytochrome c, the effect is solvent dependent: we find a decrease in ka with voltage for water but no appreciable change for HEPES. We also observe the linear coefficient to the density expansion of the cavity function (as introduced in Eq. 11), obtained from the slope of the linear region, to decrease (in magnitude) with voltage for albumin in both solvents but to remain roughly unchanged for cytochrome c.

1.5 Discussion We observe an applied potential to have a significant effect on adsorption beyond the transport-limited regime. The apparent adsorption rate constant introduced in and around Eq. 11 – a measure of the rate of adsorption to an empty surface and a physical property depending on protein/ surface/ solvent characteristics – increases with applied voltage for the negatively charged albumin, decreases somewhat for the positively charged cytochrome c in water, and remains roughly constant for cytochrome c in HEPES. Interestingly, we observe the rate of adsorption onto

18

P.R. Van Tassel

surfaces of higher coverage to increase significantly with voltage in all systems investigated. Since these measurements occur well beyond the transport-limited regime, this reflects an increased rate of attachment among protein molecules at or near the surface. Since direct alteration of the apparent rate constant by interactions with previously adsorbed proteins is unlikely, the increase in overall adsorption rate is probably due to an increased amount of surface available for adsorption and, at higher coverage, a formation of multilayers. A measure of the available surface is given by the cavity function Φ (as introduced in Eq. 11), and contributing factors to its increase are: (1) the adsorption of protein molecules with smaller projected areas, as would occur in an end-on orientation of a nonspherical protein, and (2) the formation of protein clusters on the surface. We quantify changes in Φ with voltage by considering its first-order density expansion coefficient, C1 , obtained from the slopes of the linear regions of dΓ/ dt versus Γ curves. As reported in Table 2, we observe C1 to decrease with voltage for albumin, but to be roughly independent of voltage for cytochrome c (see Table 2). A smaller adsorbed molecule projected area, or adsorption directly to the edge of growing clusters, are physical explanations consistent with this observation. Both of these features are engendered by adsorption of protein oriented such that a complementary (in this case negative) charge faces the adsorbing substrate. In this way, even the positively charged cytochrome c may adsorb more rapidly with applied voltage. Assuming a rectangular array of adsorbed proteins, the dimensions of which are the same as those previously reported [8 × 8 × 3 nm for albumin (He and Carter 1992), 3.7 × 2.5 × 2.5 nm for cytochrome c (Lvov et al. 1995)], the maximal densities are approximately 0.46 and 0.33 µg/cm2 for albumin and cytochrome c, respectively. Clearly, multilayer adsorption under applied voltage is occurring for both proteins in water and possibly, following sufficient time, for cytochrome c in HEPES. As mentioned above, adsorption in the first layer probably involves attachment at a region of complementary charge. This in turn may expose regions of opposite charge, to which charged patches of incoming proteins may be attracted. A plausible cause of multilayer adsorption is therefore the axial interaction between aligned dipoles. One expects this interaction to become weaker with each layer, however, due to incomplete charge reversal. Thus, even systems exhibiting multilayers possess kinetic curves with negative curvature, if not true plateaus. We observe the overall extent of adsorption and the influence on adsorption by an applied voltage to be greatest when employing a water solvent. One issue here is solvent quality: HEPES is a better solvent, as evidenced by higher protein solubility. Another issue is the higher pH of HEPES (7.4) compared to deionized water (5.5–6.0). For a given potential difference, this


19

works to lower both ITO and platinum potentials (as shown in Fig. 4a,b). However, when comparing adsorption at roughly equal electrode potentials (e. g., water at ∆V = 1.0 V and HEPES at ∆V = 1.5 V), we continue to observe a greater degree of adsorption in the water system. A third issue is ionic strength: the ability of charged species in solution to screen electrostatic interactions is well established and this is clearly serving to reduce in magnitude the effective (or “zeta”) potential of the proteins and surfaces investigated here. For example, screening of dipolar interactions by ionic species in the HEPES solvent seems to suppress multilayer formation. An applied potential clearly offers an opportunity to influence the structure and formation kinetics of an adsorbed protein layer. Interestingly, our work, and that of others (Asanov et al. 1997, 1998; Bernabeu and Caprani 1990; Bos et al. 1994; Feng and Andrade 1994; Fievet et al. 1998; Fraaije et al. 1990; Khan and Wernet 1997) shows that observed behavior does not always follow that expected from basic electrostatic considerations. There is little agreement on the cause of such behavior and, therefore, on the very nature of an applied potential’s influence on protein adsorption. Proposed explanations invoke surface-bound counterions, local pH effects, interfacial solvent structure, and protein charge heterogeneity. We will now comment upon each of these effects:

1.5.1 Surface-Bound Counterions Could similarly charged surfaces be attracted to one another given the presence of a third, oppositely charged species? At first glance this would seem unlikely since counterions would at most completely compensate the charge, leaving two neutral surfaces. The classic Poisson-Boltzmann approach, in which solvated electrolytic species are assumed uncorrelated, predicts only repulsion between like charged objects in an electrolytic solution. However, there are examples where charge overcompensation may occur. A spectacular recent example of this is the layer-by-layer method of growing thin films by exposing a substrate alternatively to solutions of polyanions and polycations (Decher and Schlenoff 2003). Although not fully understood theoretically, the idea is that for entropic reasons, charged sites on the polyions exchange nearly completely with smaller ions at the surface. However, in addition to sites contacting the surface, the polyions generally contain additional charged sites that extend away from the surface, thus resulting in charge overcompensation. However, this type of charge overcompensation would not be expected from the relatively small ionic species present in most protein adsorption systems.

20

P.R. Van Tassel

1.5.2 Local pH Effects Proteins are amphoteric, that is, they possess a number of acid/base sites whose charge is pH dependent. Since a positively (negatively) charged surface will repel solvated protons (hydroxide ions) and thus raise (lower) the pH near to the surface, the sign of a protein’s overall change could change upon approach to the surface. In this case, attraction that is indeed electrostatic in origin may follow. Could this effect explain the antielectrostatic behavior reported in the literature? Evidence to suggest otherwise lies in the following observation: the degree of buffering has been shown to have little effect on these antielectrostatic adsorption trends (Brusatori and Van Tassel 2003).

1.5.3 Solvent Interfacial Structure Solvent molecules (typically water) at or near a solid substrate often experience a hindered rotation relative to those far from the interfacial region. If liberated from the surface via, for example, displacement by an adsorbed protein, a net entropy gain results. An estimation of the rotational contribution to the entropy is given by the statistical mechanical result for an ideal gas rigid rotor, Srot = Nk ln qrot where N is the number of molecules, k is the Boltzmann constant, and qrot is the rotational partition function. For water at 300 K, qrot ≈ 30, so the contribution per mole of water liberated from the surface is about 28 J/Kmol. An average-sized protein might result in the liberation of 50 water molecules, so the contribution to the overall free energy of adsorption at 300 K from the solvent rotation would be about –420 kJ/mol. Of course, this is an upper limit since the molecules at the surface do possess some rotational freedom. This estimate would most accurately apply to water at a neutral, hydrophobic surface, where the molecules are expected to be very ordered. As the magnitude of surface charge is increased, the water becomes more disordered (i. e., they may experience enhanced rotation), and the free energy gain from its liberation is thus diminished. As several of the aforementioned studies show trends consistent with these thoughts, the role of solvent structure in the overall influence of an applied electric field cannot be discounted.

1.5.4 Protein Charge Heterogeneity Based on electrostatics, can a positively (negatively) charged protein be attracted to a positively (negatively) charged surface? Within the Poisson-


21

Boltzmann approach, the answer is yes, according to Asthagiri and Lenhoff (1997), so long as both protein and surface contain charge heterogeneity. The idea is that regions of complementary charge may come into close contact, while overall repulsion is sufficiently screened by counterions so that a net attraction occurs. This explanation appears to most plausibly explain our observed increase in cytochrome c adsorption with increasing adsorbent potential.

1.6 Conclusions An electric potential applied to an adsorbing substrate offers the possibility of controlling the structure of an adsorbed protein layer. This represents a powerful method of producing nanostructured coatings for biosensing, tissue engineering, enzymatic catalysis, bioseparation, and bioelectronics applications. Only in the past few years have techniques become available to measure the adsorbed layer properties in situ during formation. While several puzzling observations remain to be resolved, and while quantitative prediction of the effects of the applied potential, such as counterion binding, local pH enhancement, solvent structure, and charge heterogeneity remains sparse, the prospects are strong for a general understanding, in the not-to-distant future, of the influence of substrate electric potential on protein adsorption, and the engineering of protein coatings to nearexact specifications using voltage-based methods is certainly a realistic and laudable goal.

References Asanov AN, DeLucas LJ, Oldham PB, Wilson WW (1997) Heteroenergetics of bovine serum albumin adsorption from good solvents related to crystallization conditions. J Colloid Interface Sci 191:222–235 Asanov AN, Wilson WW, Odham PB (1998) Regenerable biosensor platform: a total internal reflection fluorescence cell with electrochemical control. Anal Chem 70:1156–1163 Asthagiri D, Lenhoff AM (1997) Influence of structural details in modeling electrostatically driven protein adsorption. Langmuir 13:6761–6768 Bernabeu P, Caprani A (1990) Influence of surface charge on adsorption of fibrinogen and or albumin on a rotating disk electrode of platinum and carbon. Biomaterials 11:258–264 Bos MA, Shervani Z, Anusiem ACI, Giesbers M, Norde W, Kleijn JM (1994) Influence of the electric potential of the interface on the adsorption of proteins. Colloids Surf B 3:91–100 Brusatori MA, Tie Y, Van Tassel PR (2003) Protein adsorption kinetics under an applied electric field: an optical waveguide lightmode spectroscopy study. Langmuir 19:5089– 5097 Brusatori MA, Van Tassel PR (2003) Biosensing under an applied voltage using optical waveguide lightmode spectroscopy. Biosens Bioelectron 18:1269–1277

22

P.R. Van Tassel

Calonder C, Tie Y, Van Tassel PR (2001) History dependence of protein adsorption kinetics. Proc Natl Acad Sci U S A 98:10664–10669 Calonder C, Van Tassel PR (2001) Kinetic regimes of protein adsorption. Langmuir 17:4392– 4395 Carignano MA, Szleifer I (2002) Adsorption of model charged proteins on charged surfaces with grafted polymers. Mol Phys 100:2993–3003 de Feijter JA, Benjamins J, Veer FA (1978) Ellipsometry as a tool to study the adsorption of synthetic and biopolymers at the air–water interface. Biopolymers 17:1759–1772 Decher G, Schlenoff JB (eds) (2003). Multilayer Thin Films. Weinheim, Wiley-VCH Fang F, Szleifer I (2003) Competitive adsorption in model charged protein mixtures: equilibrium isotherms and kinetics behavior. J Chem Phys 119:1053–1065 Feng L, Andrade JD (1994) Protein adsorption on low-temperature isotropic carbon 2. Effects of surface charge of solids. J Colloid Interface Sci 166:419–426 Fievet P, Mullet M, Reggiani JC, Pagetti J (1998) Influence of surface charge on adsorption of a hydrophobic peptide onto a carbon surface by capacitance measurements. Colloids Surf A 144:35–42 Fraaije J, Kleijn JM, Vandergraaf M, Dijt JC (1990) Orientation of adsorbed cytochrome-c as a function of the electrical potential of the interface studied by total internal reflection fluorescence. Biophys J 57:965–975 Haynes CA, Norde W (1994) Globular proteins at solid/liquid interfaces. Colloids Surf B 2:517–566 He XM, Carter DC (1992) Atomic structure and chemistry of human serum albumin. Nature 358:209–215 Khan GF, Wernet W (1997) Adsorption of proteins on electro-conductive polymer films. Thin Solid Films 300:265–271 Lvov Y, Ariga K, Ichinose I, Kunitake T (1995) Assembly of multicomponent protein films by means of electrostatic layer-by-layer adsorption. J Am Chem Soc 117:6117–6123 Malmsten M (ed) (1998a) Biopolymers at Interfaces. Surfactant Science Series. New York, Marcel Dekker Malmsten M (1998b) Formation of adsorbed protein layers. J Colloid Interface Sci 207:186– 199 Oberholzer MR, Lenhoff AM (1999) Protein adsorption isotherms through colloidal energetics. Langmuir 15:3905–3914 Ramsden JJ (1993) Review of new experimental techniques for investigating random sequential adsorption. J Stat Phys 73:853–877 Ravichandran S, Talbot J (2000) Mobility of adsorbed proteins: a Brownian dynamics study. Biophys J 78:110–120 Roth CM, Lenhoff AM (1993) Electrostatic and Vanderwaals contributions to protein adsorption – computation of equilibrium constants. Langmuir 9:962–972 Roth CM, Lenhoff AM (1995) Electrostatic and Van-Der-Waals contributions to protein adsorption – comparison of theory and experiment. Langmuir 11:3500–3509 Stankowski S, Ramsden JJ (2002) Voltage-dependent coupling of light into ITO-covered waveguides. J Phys D Appl Phys 35:299–302 Tie Y, Calonder C, Van Tassel PR (2003) Protein adsorption: kinetics and history dependence. J Colloid Interface Sci 268:1–11 Tiefenthaler K, Lukosz W (1989) Sensitivity of grating couplers as integrated optical chemical sensors. J Opt Soc Am B Opt Phys 6:209–220 Van Tassel PR (2003) Biomolecules at interfaces. In: Mark HF (ed) Encyclopedia of Polymer Science and Technology, 3rd edn. Wiley Interscience, New York, pp 285–305 Voros J, Ramsden JJ, Csucs G, Szendro I, De Paul SM, Textor M, Spencer ND (2002) Optical grating coupler biosensors. Biomaterials 23:3699–3710

2

From Kinetics to Structure: High Resolution Molecular Microscopy Jeremy J. Ramsden

Abstract. Existing methods of protein structural determination either present a static picture under unnatural conditions or impose a considerable perturbation on the molecule during measurement. This chapter introduces high resolution optical waveguide lightmode spectroscopy (OWLS) as an alternative non-imaging technique of great power and versatility enabling the structure and dynamics of proteins at interfaces to be characterized with unprecedented richness and precision.

2.1 Introduction Our knowledge of the extraordinary versatility of proteins has been greatly enhanced by the combination of new experimental techniques capable of revealing structural and dynamical details with ever deeper understanding of the molecular mechanisms of processes within living organisms. Hence, “structure” can often now be correlated with “function”. While these two terms are not very easy to define in a fully satisfactory way, it can be done ostensively, an excellent example being provided by the story of haemoglobin. The peculiar features of its function, most notably its sigmoidal binding affinity for oxygen, were already known and molecular models inferred well before the details of its structure were discovered through X-ray diffraction (XRD) (e. g. Perutz 1993). Although XRD has since become the workhorse of protein structure determination, it suffers from two significant weaknesses: 1. It presents an essentially static picture of a single structure1 2. Proteins must be present as crystals, and the exigencies of crystallization require that the proteins be placed in unusual salt conditions (good “salting out” salts (Cacace et al. 1997)). Jeremy J. Ramsden: Department of Materials, Cranfield University, MK43 0AL, UK, E-mail: j.ramsden@cranfield.ac.uk 1 Partly mitigated by analysing the Debye–Waller factors of the diffraction pattern. Principles and Practice Proteins at Solid–Liquid Interfaces Philippe Déjardin (Ed.) © Springer-Verlag Berlin Heidelberg 2006

24

J.J. Ramsden

Only very recently has the availability of intense synchrotron sources enabled smaller and less perfect crystals, hence growable under less exigent conditions, to be measured, and has also permitted snapshots and rapid sequences of snapshots to be taken, providing time-resolved structural information. But this work is still quite preliminary. Alternative approaches have, therefore, been tirelessly sought, in which proteins could be observed in their native environments and sequences of structural changes resolved temporally. One of the most promising of the new approaches has been singlemolecule observation techniques developed in a number of laboratories, mainly in Japan, in which a label (e. g. a fluorophore) is attached to the protein, and high resolution optical microscopy used to observe its motions. This impressive work has led to very significant new insights into the working mechanism of molecular motors being obtained (Yanagida et al. 1993, Funatsu et al. 1995, Ishijima et al. 1998). Note that many of the problems investigated in this field are connected with proteins at some kind of interface, most typically the solid/liquid interface, where the solid part may be a bilayer liquid membrane, and the liquid an aqueous fluid mimicking the cytoplasm. Much interest has also been shown in the atomic force microscope (AFM), but it seems fair to say that its most successful application has been for the imaging of regular arrays, to which averaging and refinement can be applied to achieve truly molecular resolution (as it is of course with XRD). Although there are examples showing how AFM of individual proteins can provide important corroborative information, there is not the same strong line of inference from the raw measured data to the desired molecularly resolved information as exists for XRD; analysis often boils down to the subjective assessment of images, and high resolution kinetics is problematic. With both these techniques (single-molecule observation and AFM), although the protein is measured in a native environment, the molecule is severely perturbed. Although the label required in the fluorescent technique may be small relative to the entire protein molecule (and even this is often not the case), it is likely to be large relative to some functionally important part. With AFM, the contact between the cantilever-mounted tip and the protein is akin to using an elephant to move a football—irreparable damage to the latter is likely to result. Hence the search for non-invasive, native-compatible and dynamically sensitive techniques continues. Note that both the single-molecule observation and AFM techniques require that the protein under investigation is present at a solid/liquid interface. Since the majority of proteins in their native state operate at such an interface, most commonly the surface of a bilayer lipid membrane, this is not a restricting condition.

2 From Kinetics to Structure: High Resolution Molecular Microscopy

25

The purpose of this chapter is to show that another new technique, well adapted to the study of proteins at or near interfaces, namely high resolution optical waveguide lightmode spectroscopy (OWLS), is capable of providing structural information superior to that obtainable from AFM, with excellent time resolution, in native environments, and without the need for labels. It is well known that even though the “ruler” of optical methods is of the order of a wavelength, i. e. hundreds of nanometres for visible radiation, interferometry can resolve structural changes with subnanometre resolution and nanosecond time resolution (Rembe et al. 2005). In the same way, OWLS, also known as high resolution molecular microscopy (HRMM), can also achieve subnanometre resolution of key average characteristic lengths of molecular systems. The basic experimental arrangement is to place the interfacial region under examination on a planar optical waveguide. For example, if the interaction of a protein with lipid bilayers is being studied, the waveguide is coated with the bilayer, which is placed in contact with a dilute solution of the protein. The structure of the remainder of this chapter is as follows: The second section focuses on OWLS, describing the fundamentals of reflectance-based methods for interrogating objects at an interface in general and OWLS in particular (briefly contrasting it with other optical methods); the next section outlines the practical determination of the lightmode spectrum; the section on static structures examines how static structural parameters of adlayers are obtained; the following section discusses kinetic analysis and dynamic structural inference; and the last section discusses advanced topic in kinetic analysis and dynamic structural inference.

2.2 Optical Waveguide Lightmode Spectroscopy In the latter half of the 17th century Newton (1730) demonstrated the existence of an evanescent field when light falls on the interface between dense and rare optical media and is totally reflected back into the dense medium. During the intervening years the phenomenon has been extensively studied, and has given rise to important measurement techniques for investigating interfacial phenomena (cf. Table 1). The overall approach is for the optically dense medium (typically a transparent dielectric) to form the substrate in contact with the optically rare medium (typically an aqueous solution) from which particles accumulate at the interface between the two media. Any change of electronic polarizability within the evanescent field, which decays exponentially into the rare medium, gives rise to a change of phase of the totally internally reflected light, which is essentially

26

J.J. Ramsden

Table 1. Theoretical foundations of different optical techniques for investigating interfacial phenomena and their experimental development, with the names of the pioneering contributors and of the instrument makers. Acronyms: SAR, Scanning Angle Reflectometry; SPR, Surface Plasmon Resonance. SAR does not make use of the evanescent field as such. In SPR, the field is generated by the collective vibrations of electrons in a thin metal film deposited at the interface Technique

Theoretical basis

Experimental demonstration

Laboratory instruments

Large-scale commercialization

SAR

Fresnel (1818) Bousquet (1957)

Schaaf et al. (1985)

–

Not anticipated

Jamin (1850)

Rudolf (1976)

Not anticipated

Ellipsometry Fresnel (1818) van Ryn van Alkemade (1883) SPR Maxwell-Garnett (1906) OWLS Tien (1977)

Turbadar (1959) Biacore (1990) Anticipated? Ramsden (1993a)

MicroVacuum Corning (2006) (2001)

the parameter measured. Table 1 provides a brief historical survey of the relevant developments up to the present. Ample comparison of these different techniques has been made elsewhere (e. g. Ramsden 1994, 1997); the next table summarizes some advantages and disadvantages. We shall take a further comparative look in Sect. 2.4. Table 2. Brief summary of some of the pertinent advantages and disadvantages of different optical techniques. Regarding the sensitivity, OWLS is intrinsically about 10 times more sensitive than SPR (for the determination of a single parameter) (Lukosz 1991). SPR only occurs with the p polarization; practically, SAR also has that restriction. See the chapter in this volume by Poksinski and Arwin concerning ellipsometry Technique

Advantages

Disadvantages

SAR

Comprehensive theoretical analysis

Slow Not very sensitive

Ellipsometry Very versatile Extensive literature

No closed form solutions of equations Not very sensitive

SPR

Strong commercial development Not very sensitive Requires a metal substrate

No closed form solutions of equations

OWLS

Highly informative closed form solutions of equations Very sensitive; versatile

Cannot be used with opaque substrates Compatible with other photonics devices


27

2.2.1 Principles of Optical Biosensing Fresnel’s coefficients of reflection give the ratio of reflected to incident amplitude for beams incident from the high refractive index side (labelled F) onto an interface between two transparent dielectrics F and J. Their refractive indices are nF and nJ respectively, with nF > nJ 2,3 . The coefficients are: Rρ =

2ρ

2ρ

2ρ

2ρ

kF / nF − kJ / nJ kF / nF + kJ / nJ

(1)

for the perpendicular (s, ρ = 0) and parallel (p, ρ = 1) polarizations, with kF = knF cos θ

(2)

where k = 2π/λ, the wavenumber for the light in vacuo, θ is the angle of incidence measured from the normal to the interface, and (3) kJ = k n2J − n2F sin2 θ . For a thin film A of thickness dA interposed at the interface, these expressions are modified by summing the reflections and transmissions at the two interfaces, yielding (Bousquet 1957): RF,A,C =

RF,A + RA,C e−2iβA 1 + RF,A RA,C e−2iβA

(4)

where βA is the phase thickness of A, defined by

βA = dA | kA |

(5)

with kA defined by Eq. 3 with J = A. These expressions may be straightforwardly extended to multiple thin layers in an interfacial stack (Ramsden 1993a). Careful measurement of the amplitude of reflected light as a function of incident angle can, in principle, yield enough data to enable Eq. 1 to be fitted to the data with the unknown parameters (which are typically the opto-geometric parameters of the thin interfacial film) as fitting variables. Often the measurement is carried out with p-polarized light in the vicinity 2

In general the media may also absorb some light, i. e. nF and nJ are actually complex numbers, but typically the imaginary part is very small compared to the real part, and for the convenience of writing we shall suppose it to vanish. 3 Horváth et al. (2002) have demonstrated reverse symmetry waveguide sensing with n < n . F J

28

J.J. Ramsden

of the Brewster angle for the F,S interface (at which the reflected component vanishes). In ellipsometry, it is the change in polarization upon reflection or transmission which is measured. The phenomenon of total internal reflection, which occurs when θ exceeds a critical angle (such that the angle of transmission would become imaginary) was known to Newton (1730), and may indeed have been discovered by him (at any rate he gives the first account published in writing). In effect, light propagating in a planar optical waveguide, i. e. a thin slab of high refractive index material sandwiched between slabs of low index material, does so via alternate total internal reflections at the two boundaries in a zig-zag path, which of course implies penetration of the light into the optically rarer medium beyond the waveguide, and is equivalent to a phase shift Φ (Eq. 8). The guided mode within the waveguide is a stationary wave, and beyond the waveguide an exponentially decaying evanescent field (Ramsden 1993a). The point at which the light turns back toward the waveguide exactly equals the characteristic decay length of the evanescent field.

2.2.2 Mode Equations for OWLS The mode equations, which govern the relation between the effective refractive indices N and the optogeometric parameters of the waveguide, can be easily derived from Eqs. (1–3), bearing in mind that R, being in general complex, can be written as the product of its modulus and argument: ˆ =| R ˆ | eiΦ R

(6)

and by noting that the existence criterion for a guided mode is that the sum of the phase shifts occurring at the two reflections and upon twice traversing the width of the waveguide must sum to an integral multiple of 2π, otherwise destructive interference occurs (Tiefenthaler and Lukosz):

ΦS,F + 2βF + ΦF,A,C = 2πm , m = 0, 1, 2, ...

(7)

where m is the mode number, and the phase changes upon reflectance are derived from Eqs. 1 and 6: 1/ 2 nF ρ N 2 − n2S ΦF,S = −2 arctan . (8) nS n2F − N 2 This assumes that the F,S interface is flat and abrupt.


29

Introducing the normalized propagation constant b (Ghatak and Thyagarajan 1989): b = (N 2 − n2S )/ (n2F − n2S ) ,

(9)

the asymmetry parameter a a = (n2S − n2C )/ (n2F − n2S ) ,

(10)

and the dimensionless waveguide parameter W: W = kdF (n2F − n2S ) ,

(11)

where k = 2π/λ is the wavenumber of the light in vacuo with wavelength λ, the mode equations for a three layer waveguide (S,F,C) are then

b 1/ 2 nF 2ρ b+a 1/ 2 2ρ + nC 1−b 1−b nF tan[W(1 − b)1/ 2 ] = (12) 1 / 2 nS 1 − [b(b + a)] (nS nC / n2F )−2ρ(1−b) with ρ = 0 and 1 for the TE (transverse electric, i. e. s-polarized) and TM (transverse magnetic, i. e. p-polarized) modes, respectively. The lower cutoff is given by 2ρ n F Wc(ρ) = arctan a1/ 2 + mπ (13) nC and the penetration depth by

∆z = 1/ [k(N 2 − n2C )]1/2 .

(14)

ΦF,A,C can be calculated if the adlayer A is assumed to be a uniform, homogeneous film (Tiefenthaler and Lukosz 1989). Measurement of two modes (typically the zeroth order TM and TE) and the simultaneous solution of the corresponding two mode equations enables the thickness and refractive index of the adlayer to be calculated. The explicit solutions for nA and dA are given by Guemouri et al. (2000). As discussed in the next subsection, these two parameters can then be combined to yield the number of particles ν captured at unit area of interface according to ν = dA

nA − nC dn/ dc

(15)

where dn/ dc is the refractive index increment of the particle, dependent on its polarizability and the medium in which it is dissolved or suspended (Ball and Ramsden 1998)

30

J.J. Ramsden

If the adlayer is not isotropic, then it must be characterized by two or more refractive indices. A fairly common case is for the adlayer to be uniaxial, and hence characterized by two refractive indices, no and ne (ordinary and extraordinary). They are related to the measured effective refractive indices by no = nA,TE

(16)

where nA,TE is the adlayer refractive index in the TE mode equation (Eq. 7), and −1/ 2 1 nF 2 1 1 ne = + − . (17) n2o NTM n2A,TM n2o In solving the mode equations, if there are only two of them (e. g. corresponding to the TE0 and TM0 modes), then dA must be measured independently in order to proceed. On the other hand, if thicker waveguides are used, supporting higher modes, then all the parameters may be determined independently (albeit at lower sensitivity (Tiefenthaler and Lukosz 1989, Mann 2001).

2.2.3 The Uniform Thin Film Approximation (UTFA) For many of the applications relevant to this volume, nanoscale particles such as proteins are added to the interface, and the question arises whether the resulting monolayer deposit, which is inherently heterogeneous, can be modelled as a homogeneous, uniform thin film. Were that to be the case the analysis of the raw data would be much easier than otherwise. Careful comparison of results assuming such a film with explicit consideration (using Mie theory) of the actual heterogeneity of such layers has revealed that the UTFA is an excellent approximation up to particle diameters of around 50 nm (Mann et al. 1997). These films are often isotropic, or anisotropic to a negligible degree, which means that they are fully characterized by just two parameters, refractive index nA and geometrical thickness dA 4 . In OWLS, these two parameters can be obtained by a straightforward analytical solution of two simultaneous mode equations using the effective refractive indices of two independent modes as the input. A disadvantage of 4 One way in which the assumptions of the UTFA could be invalidated even though minute nanoparticles are being adsorbed as a monolayer is if those particles cluster together to create large islands. Such behaviour is evidenced by a broadening of the TM incoupling peak but not the TE one (J.J. Ramsden and T. Jaworek, unpublished observations), and has a definite kinetic signature as discussed in Sect. 2.6.1.


31

the other optical methods (viz., SAR, ellipsometry and SPR) is that tedious fitting calculations are required to extract nA and dA . While the general availability of powerful personal computers makes this less of a problem than formerly, the fitting procedure is itself not robust insofar as a large family of possible solutions satisfy the equations. In many applications, however, only the number ν of particles per unit area of interface (or the amount M of protein deposited on a membrane) is required, and this can be straightforwardly obtained from any nA and dA pair yielding a solution5 by ν = M/ m = dA (nA − nC )/ (m dn/ dc)

(18)

where m is the mass of one adsorbate particle (e. g. a single protein molecule) and dn/ dc is the refractive index increment (rii) of the adsorbate solution. The rii simply expresses the slope of a plot of solution refractive index versus the concentration of protein which it contains. For a given protein dissolved in a given solvent, this has been shown to be linear up to concentrations of around 0.5 g/ml, in which range even dense monolayers comfortably fall (de Feijter et al. 1978). Some common and other blood proteins dissolved in common buffers used in biological research all have a value of about 0.18 cm3 /g (de Feijter et al. 1978, Ball and Ramsden 1998).

2.2.4 Optical Invariants A weakness of the method outlined in Sect. 2.2 for evaluating OWLS data is that the results depend in an ill-defined fashion on the model chosen to describe the adlayer A. For many adlayers encountered in practice, especially those formed from macromolecules captured at the solid–liquid interface, the assumption of a uniform, homogeneous film is unrealistic. The use of optical invariants enables key parameters characterizing the adlayer to be calculated without the need for possibly unwarranted assumptions. The difference between the optical responses of the idealized Fresnel interface and the real interface is given in terms of surface excess polarization densities (Mann 2001) (cf. Gibbs’ surface excess quantities in his treatment of the thermodynamics of thin films). Any measurable quantities should be independent of the position of the (fictitious) Fresnel interface: such optical invariants are obtained by combining the polarizabilities obtained by a multipole expansion of the surface excess polarization densities 5

More rigorously, this implies that nA and dA are not optical invariants of the system. The optical invariants arise analogously to the Gibbs formulation of interfacial thermodynamic properties: they are such quantities that are invariant under small movements of the interfacial boundary (see Sect. 2.2.4).

32

J.J. Ramsden

such that the combinations are invariant with respect to displacement of the idealized Fresnel interface. These optical invariants allow determination of the maximum information available from the data with the minimum of ambiguity. Mann (2001) has derived the optical invariants for OWLS and given analytic equations relating the parameters obtained from OWLS measurements to the optogeometric parameters of uniaxially anisotropic adlayers. The first order non-invariant excess polarization density parallel to the interface is

γ˜e = kdA n˜ 2

(19)

where n˜ is the excess parallel refractive index, i. e. n˜ 2 = n2 − n2C

(20)

where n is the refractive index parallel to the interface, i. e. the ordinary refractive index, and we can call n˜ 2 / n2C the increment of the relative dielectric constant of the adlayer. The first and second order (in dA /λ) invariants are J1 , J21 , J22 and J23 : J1 =

n2 − n2F − (n2F − n˜ 2 )α n2C / n˜ 2 + 1 + α

kdA

(21)

where the optical anisotropy of the adlayer is

α=

n2⊥ − n2 n˜ 2

,

(22)

where n⊥ is the refractive index perpendicular to the interface, i. e. the extraordinary refractive index; 2 n˜ n2 J21 = C − 1 (kdA )2 , (23) 2 NFC where NFC = n2F − n2C ,

(24)

i. e. depending only on polarization density parallel to the interface; J22 = −

1 − n˜ 2 (1 + α)/ NFC + α/ 2 (kdA )2 ; n2C / n˜ 2 + 1 + α

(25)

and J23 = J21 − J22

n2F n2C . n2F + n2C

(26)


33

Table 3. Comparison of different techniques in terms of obtainable quantities Technique

NonOptical invariants invariant 1st ordera 2nd ordera

OWLS γ˜e Ellipsometry SAR a

J1 J1 J1

J22 , J23 J23 J22 , J23

in dA /λ

The power of the optical waveguide technique is that four opto-geometric parameters characterizing an (anisotropic) adlayer can be extracted. Table 3 compares the available parameters with those from other optical techniques. If dA /λ is small (< 0.01), then γ˜e and J1 dominate the waveguide optical response, and (Mann 2001) 2 2 1/ 2 − n N NFC C γ˜e = − 1/2 πm − kdF NFE1/2 + 2 arctan TE (27) NFE NFE where 2 NFE = n2F − NTE

and N 2 − NCF J1 = 2TM NTM NCF / n2F

(28)

NFC 1/ 2 γ˜e − 1/2 πm − kdF NFM NFM 1/ 2 2 − n2C n2F NTM + 2 arctan , NFM n2C

(29)

where NCF =

n2F n2C n2F + n2C

(30)

and 2 . NFM = n2F − NTM

(31)

The adlayer mass M per unit area, for thin films, i. e. dA /λ < 0.01, is given by M ≈ −˜γe (dc/ dn2 )/ k , i. e. depending only on the TE modes.

(32)

34

J.J. Ramsden

2.3 The Practical Determination of Waveguide Parameters The fundamental parameter of the optical waveguide is the effective refractive index N. Equivalently, one may refer to the phase velocity, propagation constant, reflection phase change, etc.—all are equivalent (Ramsden 1993a). N characterizes a guided mode. Modes are discrete, and for the thinnest waveguides, which are the most sensitive to changes in the interfacial region, are well separated. It is important to realize that even though the overall waveguiding structure through which light is propagating comprises several layers, each with a distinct refractive index, the light propagates through all of them with a common phase velocity. The two main methods for determining the lightmode spectrum are (i) coupling, and (ii) interferometry. In this chapter we shall only deal with the former, for it most readily and directly allows the precise and accurate determination of the absolute values of the effective refractive indices of two or more modes, which are required for the structural analysis that is our ultimate goal. Interferometry is covered in the references cited earlier (Ramsden 1994, 1997), and see also the chapter by Freeman in this volume. Coupling requires a grating (or, less conveniently, a prism). Consider a slab structure consisting of a support layer S (typically made from optical glass about a few tenths of a mm thick), a thin high refractive index transparent film F (typically a metal oxide ∼ 100 nm thick), and the cover medium C (typically air or water). An optical grating is incorporated into the waveguide at the F,S or F,C interface. A shallow grating, a few nm deep, is sufficient for coupling the few percent of light needed for measurement purposes. In a typical incoupling configuration, an external beam impinges onto the grating, making an angle α with the grating normal. The wavenumber component in the direction of guided propagation is then nair sin α + 2π/Λ, where = 0, ±1, ±2, ... is the diffraction order and Λ the grating constant. If this matches a guided mode with wavenumber kN, incoupling will occur according to the coupling condition N = nair sin α + λ/Λ .

(33)

The procedure is, therefore, to measure the light emerging from the end of the waveguide while varying α. The emerging light will appear as a series of sharp peaks, successively TMm=0,=1 ,TEm=0,=1 , etc., called the mode spectrum. At the time of writing, the highest precision in the determination of N is achievable by mechanical goniometry in a temperature-controlled environment, with which α can typically be determined to microradian precision. The refractive index of most substances varies significantly with temperature. The variation is directly related to the coefficient of thermal


35

Table 4. Contributions to the uncertainty of N in an input grating coupler. It should also be noted that since ∂N /∂nC ∼ 0.1 and ∂nC /∂T ∼ 10−4 K−1 for water, aqueous solutions should be thermostated to ±0.1 ◦ C to keep within the bounds of uncertainty imposed by the other terms Parameter nair

Λ α λ

Typical value

Uncertainty

Physical origin

1.0002673 416.147 nm 0.09 rad 632.816 nm

10−7

Temperature fluctuations (T ± 1 ◦ C) Grating lateral thermal movement Goniometer mechanical instability Laser mode jumping

0.001 1.25 × 10−6 rad 0.001 nm

expansion of the substance, as can be seen from the Lorentz-Lorenz equation, which connects the macroscopic index of refraction with various atomic or molecular parameters, i. e. molecular mass Mr , density ρ and molar refractivity RM : n2 = (Mr + 2ρRM )/ (Mr − ρRM ) .

(34)

Of these, only the density varies with temperature. Therefore, it is obviously important to record the temperature of the sample under investigation when making optical measurements. Table 4 lists the various sources of uncertainty in the determination of the mode spectrum peaks. After applying the usual combining laws it can be seen that all factors contribute roughly equally to the overall uncertainty in N, typically ±1 − 2 × 10−6 .

2.3.1 Device Fabrication The optical waveguides and grating couplers can be made by a variety of routes. For fabricating the thin layers of high refractive index transparent oxides, either physical vapour deposition (e. g. evaporation or sputtering) or sol-gel pyrolysis are used. The optical quality requirements are extremely stringent: the absorption coefficient should generally be less than 10−5 cm−1 , otherwise insufficient light will reach the end of the waveguide and the signal:noise ratio will be degraded. The gratings required for coupling can be made by either topographical or refractive index modulation. The former can be achieved by structuring the optical glass support (the S layer) using the conventional methods of photolithography; to obtain the fine gratings required (Λ of the order of 400 nm or less) for working with = 1, holographic exposure of photoresist is required. Coarser gratings can be used with higher order diffraction peaks but the signal:noise ratio is lower. It is also possible to modulate the surface of the F layer: however, this is less desirable since the surface (that will interact with the proteins) may

36

J.J. Ramsden

then become chemically contaminated. Shallow gratings (5−10 nm deep) are sufficient and have the great advantage that the F layer is thereby so slightly modulated in the grating region that the mode equations derived above retain their validity. A further advantage is that, from the hydrodynamic viewpoint, the surface may still be considered planar. Perfect planarity is, however, only achievable with refractive index modulation. This can be achieved by exposing the substrate to a beam of energetic nitrogen through a mask (created, as above, by photolithography via holographic exposure of photoresist). Nitrogen atoms implanted in the silica modify its refractive index. The technology is non-trivial to perfect, however, mainly because of the difficulty of removing photoresist residues hardened by exposure to the energetic nitrogen beam. Further fabrication steps concern the coating of the waveguide with whatever material is serving as adsorbent for the proteins under investigation. Lipid bilayers are conveniently deposited on the surface using the Langmuir-Blodgett technique (Ramsden 1999); alternatively a suspension of lipid vesicles can be used, which fuse with the surface to create the bilayer (Csúcs and Ramsden 1998a). In both cases, a few molecular layers of water between the inner leaflet of the bilayer and the solid support ensure that the membrane (in its liquid crystalline state) has native fluidity, i. e. the lateral diffusivity of the lipid molecules is approximately the same as in the membrane of a living cell. Polymer layers, including polyelectrolytes, can be deposited by spontaneous attachment from solution (e. g. Ramsden et al. 1995); the large number of contacts per molecule ensures irreversible attachment. Other materials can be deposited by physical or chemical vapour deposition. A layer 10 nm thick will generally be found to completely mask the underlying F layer (Kurrat et al. 1997). The only restriction is that the layer should not absorb the guided light; hence metals can only be deposited to a thickness of about 2 nm, which is often insufficient for a continuous layer to be formed; and materials (dyestuffs) with intensive absorption bands coinciding with the wavelength of the guided light should be avoided.

2.3.2 Fluid Handling Arrangements The overwhelming majority of in situ investigations into protein structure involve aqueous solutions of proteins. Hence, a way must be found to bring the solutions in contact with the adsorbing surface. A convenient way of doing this is to clamp a cuvette to the surface of the waveguide, such that it forms the bottom of the cuvette.


37

2.4 Static Structure The procedures to be followed are implicit in the treatment given in Sect. 2.2. The parameters emerging are (for waveguides supporting zeroth and first order TM and TE modes) dA , no , ne and one other (e. g. nC ); if the adlayer is isotropic a single refractive index nA suffices. Examples of an isotropic adlayer constituted from small spheroidal proteins is given by Guemouri et al. (1998); and of a continuous anistropic layer by Ramsden (1999) and Ramsden et al. (1995).

2.5 Kinetic Analysis and Dynamic Structural Inference The canonical expression for the accumulation of particles at the solid/liquid (and indeed the solid/gas) interface is dM/ dt = ka c∗ φ(M) − kd (t, M)M

(35)

where M is the mass (per unit area) of accumulated material (sometimes we shall use ν to denote the number of adsorbed particles per unit area), ka is the chemical adsorption coefficient (which may depend on ν); c∗ , the effective bulk solution concentration of the particles; φ, the available area function; and kd , the chemical desorption “coefficient” (which may depend on both M and t). In the following subsections, each of these factors will be analysed.

2.5.1 Particle Transport We first suppose that our particles are smooth, rigid and spherical and totally buoyant, i. e. their movement is not influenced by terrestrial gravity. A particle suspended in a fluid can be transported by convection (i. e. entrained by the motion of the fluid) or by diffusion. The key parameter is the relative motion of the fluid with respect to the adsorbent. Far from the surface, the flow is uninfluenced by the surface; at the surface, on the other hand, friction dictates that the fluid is stationary (ignoring the possibility of slip); and the velocity in between is constantly diminishing. Hence near a surface, a particle will move by diffusion, and far from the surface by convection. At a certain intermediate distance, typically of the order of tens of micrometres, the transport régime will cross over from convection to

38

J.J. Ramsden

diffusion. This distance is known as the diffusion boundary distance (δ). A comprehensive treatment is given by Levich (1962). The flow in certain geometries, such as the tube or plate, can be solved analytically. These geometries are therefore particularly favourable for kinetic evaluation, a factor which should be borne in mind when designing flow cuvettes. For flow in a tube, 2/ 3 DC 1/ 3 3 δ= (36) 2 F where D is the diffusivity of the protein, F the volumetric flow rate, and C a constant depending on the dimensions of the tube. When choosing flow rates, it should be remembered that only laminar flow régimes can be analysed conveniently, i. e. up to Reynolds numbers of at most around 1000. In principle a diffusion boundary still exists in the case of turbulent flow, but the motion is much more complicated than in the case of laminar flow. Hence, a protein-sized particle will be moving diffusively already at a distance of the order of a thousand times its own dimensions ( a few nm in diameter) from the adsorbing surface. At a distance of the order of ten times its own dimensions from the adsorbing surface, the particle may begin to be influenced by the long range hydrophilic repulsion (see Cacace et al. (1997) for a discussion of intermolecular forces), which will considerably retard its rate of arrival at the surface. When designing the flow conditions for an experiment, it is only necessary to ensure that convective-diffusion is rapid enough to replenish the particles lost from the solution in the vicinity of the surface by attachment to it. This is of practical importance given that many biological samples of carefully purified macromolecules are available only in extremely small quantities. Good quantitative approximations for analysing the flow régimes can be derived from the equations of Fick and Smoluchowski (see Ramsden (1998) for a more complete discussion). If the surface (at z = 0, where z is the coordinate normal to the surface) is a perfect sink for the adsorbate, then the bulk (solution) concentration cb is zero at z = 0, the concentration gradient will be approximately linear, and the rate of accumulation is dν/ dt = cb D/δ .

(37)

It is a good idea to compare the experimentally observed rate with this maximum upper limit (which may, however, be exceeded if there is a long range attractive force, e. g. electrostatic, between particle and surface). The effect of any energy barrier is to retard accumulation. In the immediate vicinity of the surface, the local bulk concentration cv will be much


39

higher than zero (although still less than cb ). It is convenient to consider that the rate of accumulation at the surface is given by the product of cv and a chemical rate coefficient ka , which is directly related to the repulsive energy barrier (Spielman and Friedlander 1974): dM/ dt = ka cv

(38)

and the Fick-Smoluchowski régime (linear concentration gradient) applies to the zone above this vicinal region. Hence

dcv (cb − cv ) V = γS D − ka cv , (39) dt δ where V and S are unit volume and surface respectively. Strictly speaking the distance of the vicinal layer from the surface should be subtracted from δ in the denominator, but since that distance is of the order of molecular dimensions, i. e. only a few nm, whereas δ is of the order of a few or a few tens of microns, this correction can be neglected. If desorption of the material also has to be taken into account a term with a chemical desorption coefficient kd can be included:

(cb − cv ) dcv =S D + kd (t)ν − ka cv . (40) V dt δ In a great many cases accumulation of material is limited to a monolayer, or to occupying a monolayer of receptors, in which case a function φ must be introduced, which gives the fraction of the surface still available for adsorption or binding (i. e. the probability that space is available). Our kinetic equation then becomes:

dcv (cb − cv ) V =S D + kd (t)ν − ka cv φ(ν) . (41) dt δ We shall discuss φ—which obviously depends on M or ν—below. One important implication is that as the surface fills up, i. e. as φ → 0, cv will tend to approach cb , and dM/ dt will asymptotically approach zero regardless of the flow régime. This can be immediately seen by letting the left hand side of whichever of the previous three equations is appropriate go to zero, yielding an explicit expression for cv , e. g. cv = cb / (1 + δφ/ D)

(42)

if ka = 1 and kd = 0, which can then be substituted into Eq. 35, in which c∗ = cv .

40

J.J. Ramsden

The pure diffusion régime If there is no flow at all then the vicinal layer is replenished by diffusion only, i. e. there is no distance at which concentration is maintained constant by effectively an infinite reservoir as in convective diffusion. This leads to the well-known result of Smoluchowski, according to which the flux to the surface constantly diminishes as t−β , where β = 0.5 for standard diffusion (other exponents have been found for the diffusion of highly non-spherical, non-compact proteins such as tenascin (Ramsden 1992). If the flow rate is slow enough for the rate of adsorption to be limited by transport alone, or, more quantitatively, if the dimensionless parameter.

ε = ka δ2 / D

(43)

becomes large relative to unity, then the initial rate of adsorption is given by Eq. 37, from which D may be obtained. This is a useful way to determine the diffusion coefficient in solution.

2.5.2 The Chemical Adsorption Coefficient If a repulsive potential barrier U(z) exists between protein and surface, where z is the distance between them, not every arriving protein will adsorb, even if there is space for it to do so, and ν˙ will be diminished by a rate coefficient ka , which can be found by integrating U(z): ⎡ ⎤ ∞ (44) ka = D ⎣ [exp(∆U(z)/ kT) − 1]dz⎦ . 0

Sometimes the denominator of the right hand term is called the “adsorption length”, δa . The interaction potential U can be approximated by the sum of the particle-surface interaction free energies, corresponding to the Lifschitz-van der Waals (LW), electrostatic (el) and electron donor-acceptor (da) interactions: U(z) = ∆G(LW) + ∆G(el) + ∆G(da) ,

(45)

with each of the terms given by

∆G(LW) = −AH r/ (6z) ,

(46)

where r is the particle radius and AH the mean Hamaker constant;

∆G(el) = 4πε0 εψ1 ψ3 ln[1 + exp(−κz)]r ,

(47)


41

where the ψ are the electrostatic surface potentials of adsorbent (particle) and adsorbate, and 1/κ is the well known Debye length; and

∆G(da) = 2πχ∆G(da) exp[(0 − z)/χ]r ,

(48)

where χ is the decay length for the da interactions, and ∆G(da) the da interaction energy for two infinite planar surfaces of the materials of adsorbent and adsorbate at “contact” (i. e. separated by a distance of approximately 0.15 nm).

2.5.3 The Analysis of The Available Area Function If we assume that the potential barrier determining ka does not change with the accumulation of material, then ka merely acts as a linear scaling factor on the adsorption rate (Eq. 35), and does not affect the form of the rate law. In contrast, the function φ affects its shape, and the main goal of many quantitative adsorption studies is to identify the function φ and correlate it with the molecular properties of the system under study. General properties of φ It is obvious that φ = 1 for a totally empty surface with θ = 0. θ is the fraction of the surface covered by adsorbed particles, and is related to ν and total adsorbed mass M by

θ = νa = Ma/ m

(49)

where a is the occupied area per particle and m its mass. The analysis of high-resolution kinetics (such as those obtainable with OWLS) allows a to be determined with an accuracy equal to or better than with atomic force microscopy. A totally full surface (θ = 1) will have φ = 0. What is less obvious is that a surface may be less than fully occupied (θ < 1) but nevertheless may not be able to offer any free place to an incoming particle. Even polygons that can tile the plane will only reach θ = 1 if they are perfectly aligned. If they are placed randomly, φ will already approach zero for θ as little as 0.5. Given that φ is a function of θ, we can quite generally expand it in powers of θ:

φ = 1 − b1 θ + b2 θ2 + b3 θ3 + ...

(50)

The so-called random sequential adsorption (RSA) has been analysed extensively in two dimensions (e. g. Evans 1993) from purely geometrical considerations. The coefficients b1 , b2 , and b3 have been determined for

42

J.J. Ramsden

Table 5. Coefficients for the available area function Mode

b1

b2

b3

Application

Langmuir

1

0

0

Clustering, large receptors

RSA (spheres)

4

3.808

1.407

Irreversible adsorption

GBDa

4(1 − j) 3.808 − 1.407 + 4.679j − 0.180j − 3.128j2 25.58j2 + 8.550j3

Nucleation and growth

j is the ratio of probabilities p/ p , p being the probability of a particle adsorbing on a hitherto unoccupied patch of the absorbent at which it arrives, and p the probability that it adsorbs after having reached a space large enough to accommodate it by a path of correlated lateral diffusion in the immediate vicinity of previously adsorbed particles. In the former case a protein arrives at an empty patch of surface, whereas in the latter case it arrives at a region where previously arrived particles are already adsorbed, and migrates to the edge of the cluster before becoming attached to the surface. j = 0, the lowest possible value, would correspond to pure random sequential adsorption (i. e. binding takes place solely independently of preadsorbed particles at empty patches of surface), and higher values correspond to increasingly favoured particle clustering. a

spheres and ellipsoids (ellipsoids with an aspect ratio of 1:4 have virtually identical coefficients to those of spheres (Viot et al. 1992), for irreversible and reversible deposition, and deposition of laterally mobile particles (reversibility and mobility only affect b3 ), etc. Experimental limitations make it unnecessary to expand φ beyond the third power of θ. Table 5 gives some values of the coefficients. Practically, by plotting the rate of adsorption against the amount adsorbed, i. e. in direct accordance with Eq. 35, it is often possible to identify the adsorption mechanism merely by visual inspection. Four principle types of behaviour are observed: 1. dν/ dt ∼ constant implies that φ = 1. This is characteristic of the formation of isolated aligned chains of adsorbate (e. g. lysozyme dissolved at low ionic strength (Ball and Ramsden (1997), with images given in Ramsden (1998)); 2. dν/ dt concave (i. e. progressively slower) implies that φ(θ) is a characteristic polynomial function of θ, implying that the proteins interact via excluded volume only, i. e. “pure” random sequential adsorption (e. g. common blood proteins such as transferrin and serum albumin (Ramsden 1993c, Kurrat et al. 1994)); 3. dν/ dt ∼ linear, implying that φ = 1 − θ (Langmuir adsorption): for adsorption onto a continuum, however, the annihilation of exclusion zones implied by this linear relation between φ and θ can only occur upon 2D clustering or crystallization (e. g. cytochrome P450 added to a membrane (Ramsden et al. 1994));


43

4. dν/ dt concave/convex implies generalized ballistic deposition (GBD) (e. g. phospholipase A2 (Csúcs and Ramsden 1998b)); 5. dν/ dt bell-shaped implies nucleation and growth of deposits (e. g. lysozyme on silica in the presence of thiocyanate ions (SCN− ) (Ball et al. 1999). Equation 35 with the appropriate expression for φ can be fitted to numerically differentiated data, or the numerically integrated version of Eq. 35 fitted to the (M, t) data, by a least-squares optimization procedure, with free parameters a, ka (and j for a GBD process, see Table 5), and kd for reversible adsorption. If there is a desorption phase, it is more robust to globally fit both adsorption and desorption to Eq. 35 with an appropriate change of boundary conditions at the moment of flooding.6 Asymptotic behaviour and jamming limit Experimental values of the jamming limit θJ (i. e. the value of θ when φ = 1) can be compared with theoretical values for spheres (Schaaf and Talbot 1989), and spherocylinders and ellipsoids (Viot et al. 1992) in order to deduce the shape of the molecule. When θ is very close to θJ , the final approach to jamming follows a power law whose exponent depends on the number of degrees of freedom of the adsorbing particle. Ellipsoids are specified by both position and orientation, i. e. they have one more degree of freedom than spheres, but the experimental data is usually too noisy to permit determination of the exponent sufficiently precisely to enable the number of degrees of freedom to be unambiguously established.

2.6 Behaviour of Real Proteins Unlike rigid colloidal particles, proteins have only limited conformational stability and any perturbation, such as adsorption to a surface, may involve energies comparable to the cohesive energy of the molecule and hence may engender conformational changes. Suppose that the enthalpy of protein adhesion to a surface equals or exceeds the enthalpy of the intramolecular contacts in the native protein conformation. Then unfolding (denaturation), which implies a gain of entropy, since many denatured conformations are possible, inevitably implies a loss of free energy, which drives the process towards adsorption with denaturation (Fernández and 6 If the particle-free solution (PFS) is introduced rapidly, then the entire diffusion boundary region will be rapidly swept clear, i. e. c(z) = 0 for all z > 0. On the other hand, if the PFS is introduced slowly, then c(z) = 0 for only z ≥ δ initially.

44

J.J. Ramsden

Ramsden 2001). In general, it can only be prevented if the surface interacts less with the protein than the protein interacts with itself, i. e. if the molecule would lose less enthalpy upon binding than upon folding into its native conformation. In considering the energy balance, it should be borne in mind that many surfaces engender changes in the solvent structure, especially of extensively hydrogen bonded solvents such as water (Wiggins 2002), and these changes may perturb the protein even before it arrives at the surface. Evidence for conformational rearrangement, including simple reorientation, can be obtained from a series of adsorption measurements carried out at different bulk concentrations. If the usual method of analysis (e. g. Ramsden 1993a) yields decreasing a with increasing cb , rearrangement can be inferred, and its characteristic time τs = 1/ ks can be obtained by equating it to the characteristic time for adsorption, τa = 1/ (Jcb φa), at the bulk concentration corresponding to the mid-point between the limiting lower and higher areas, where J is the protein flux to the empty surface normalized to unit adsorbent area and unit bulk concentration. A more sophisticated approach involves simultaneously fitting the adsorption data to equations explicitly taking rearrangement into account (Van Tassel et al. 1998): ν˙ α = ka cb φ − ks να Ψ

(51)

and ν˙ β = ks να Ψ ,

(52)

where subscripts α and β refer respectively to the lower and higher areas, and Ψ is a function analogous to φ giving the probability that space to rearrange (taking up more space) is available.

2.6.1 Evaluation of Lateral Diffusivity and 2D Crystal Unit Cell Size Extremely valuable information may be obtained by repeating adsorption measurements at different bulk concentrations. We consider three cases, A, B and C: A. In this, the simplest case, the same behaviour is obtained at all cb and the (M, t) curves may be superimposed upon each other by rescaling time as the product cb t. B. The shape of the ν˙ vs ν plot is the same for all cb , but a varies (usually increasing with decreasing cb ) (Ramsden 1993c). This implies that


45

adsorbed proteins are occupying more area after initial attachment, via orientational or conformational rearrangement. Let kr be the rate of rearrangement. The characteristic time for rearrangement τr = 1/ kr may be supposed equal to the characteristic time for adsorption τa = 1/ (Jcb φa) at the bulk concentration corresponding to the mid-point between the limiting lower and higher areas a, where J is the protein flux to the empty surface normalized to unit adsorbent area and unit bulk concentration. A more sophisticated approach involves simultaneously fitting the adsorption data to the equations explicitly taking rearrangement into account (Van Tassel et al. 1998), see Eqs. 51 and 52. C. The adsorption changes from RSA at high cb to Langmuir at low cb (Ramsden et al. 1994). At low bulk concentration, τa is long and adsorbed proteins have time to diffuse laterally, for which the characteristic time τ2 = 1/ (D2 ν), and crystallize. τ2 can be obtained from the value of cb at the crossover point when the series of ν˙ vs ν curves changes from RSA to Langmuir, at which point it may be supposed that τ2 = τa . The next step is to compare a obtained at high cb with that obtained at low cb . Near equivalence means that the proteins are forming a random cluster. If the value at low cb is much greater than at high cb , two dimensional crystallization may be inferred and the value of a at low cb corresponds to unit cell size.

2.6.2 Desorption A further extremely important extension of the basic experiment of allowing one protein to adsorb from solution at a single bulk concentration is to suddenly flood the system with protein-free solvent, which in practice generally means replacing flow of the protein solution by flow of protein-free solvent. In that case Eq. 35 becomes simply dM/ dt = −kd M .

(53)

As already pointed out, in general, kd is not a true constant. Only in the case of pure exponential decay of M can ideal, memory-free desorption be inferred, and another characteristic time, τd = 1/ kd , be defined. Usually M(t)cb =0 is strongly non-exponential. In this case it is convenient to use a memory function to characterize the adsorption behaviour. The amount of protein bound, ν(t), can be represented by the integral (Talbot 1996) t ν(t) = ka c 0

φ(t1 )Q(t, t1 )dt1 .

(54)

46

J.J. Ramsden

The memory kernel Q denotes the fraction of protein bound at epoch t1 that remains adsorbed at epoch t (if dissociation is indeed a first order (Poisson) process then Q(t) = exp(−kd t)). A necessary condition for the system to reach equilibrium is lim Q(t) = 0 .

t→∞

(55)

The dissociation coefficient is time dependent and is given by the quotient t φ(t1 )Q (t, t1 )dt1 . (56) kd (t) = 0t 0 φ(t1 )Q(t, t1 )dt1 Cooperative effects are not uncommon in densely packed protein monolayers. An example in which the desorption of an individual particle is influenced by its adsorbed neighbours is described by Kurrat et al. (1994).

2.6.3 Multilayers Many of the pure solutions of globular and fibrous proteins investigated hitherto have shown only monolayer adsorption. An exception is laminin allowed to adsorb in the presence of calcium ions (Ramsden 1993b), in which case adsorption switches from monolayer in the absence of Ca2+ to multilayer. The canonical behaviour, which is supposed to also describe the formation of amyloid plaques in higher organisms, is the sequence (starting from a dilute solution of A suddenly placed in contact with a surface): (i) adsorption of A at the solid/liquid interface; either (ii) polymerization of A (facilitated by the juxtaposition of the individual A on the surface) and conformational change of A (→ A due to polymerization) or (iii) conformational change of A (→ A due to adsorption) and polymerization of A ; (iv) adhesion of A to A . Multilayers have also been observed in the case of adsorption of complex mixtures of proteins, e. g. blood serum (Kurrat et al. 1998). The behaviour is extremely complex and not well understood. An approach starting with a mixture of the two most abundant proteins, followed by the three most abundant, etc., would doubtless help to elucidate the process, but given that there are about 1000 different proteins in blood serum, the development of high throughput screening methods capable of yielding the level of physico–chemical information currently obtainable from OWLS is required for seriously advancing towards understanding the mechanisms involved.


47

2.7 Conclusions High resolution optical waveguide lightmode spectroscopy (HR-OWLS), also known as high resolution molecular microscopy (HRMM) is able to yield detailed physico–chemical characterization of proteins, other biopolymers, polymers and other particles (even living cells (Ramsden et al. 1995)) adsorbing at and desorbing from buried surfaces such as the solid/liquid interface. The high precision of OWLS—in essence due to its very high signal/noise ratio, traceable to the very large number of multiple reflections per unit area of observed interface—enables the opto-geometric parameters of even monomolecular layers to be determined, such as the thickness (to a precision of ±10−1 nm) and the refractive indices (to a precision of about ±3×10−4 ), from which a robust determination of the predominant orientation of the molecular components of the layer can be deduced. Small conformational changes resulting from a change of conditions can be detected in this way. Much more information is obtainable if the kinetics of addition and spontaneous removal of particles to and from the surface is followed with good time resolution: the mean projected area and shape of the adsorbing particles, their lateral diffusion and sticking coefficients if they tend to cluster on the surface, the lattice parameters of two dimensional crystals if they form regular arrays, their expansion coefficient and the kinetics of expansion if they undergo area-changing reorientation or conformational change, the adsorption free energy barrier, etc. If a series of experiments under different conditions is carried out, the different contributions to the adsorption free energy barrier may be resolvable. If the memory function can be characterized, a wealth of detail concerning the dynamics of the adsorbed protein may be inferred, including cooperative effects. Additional experiments involving repeated pulses of protein solutions of different concentrations may further illuminate these processes. The HRMM approach is fundamentally different from that of scanning probe or single molecule fluorescence microscopies, its main rivals in examining the behaviour of complex nanoparticles such as proteins, in that individual objects or their locations are not imaged. Hence the radial distribution function of the deposited objects cannot be measured directly. Nevertheless, all the information obtainable from the imaging techniques can be inferred from OWLS measurements, as well as a great many essential parameters characterizing all aspects of the adsorption process in a totally non-destructive, non-invasive fashion.

48

J.J. Ramsden

Acknowledgements. The author sincerely thanks the other members of the MEMbrane-coated Optical-grating Coupler Sensors (MEMOCS) consortium for numerous valuable discussions on the subject of this Chapter during the past few years.

References Ball V, Ramsden JJ (1997) Absence of surface exclusion in the first stage of lysozyme adsorption is driven through electrostatic self-assembly. J phys Chem B 101:5465–5469 Ball V, Ramsden JJ (1998) Buffer dependence of refractive index increments of protein solutions. Biopolymers 46:489–492 Ball V, Lustig A, Ramsden JJ (1999) Lag phases in the adsorption of lysozyme to Si(Ti)O2 surfaces in the presence of sodium thiocyanate. Part I. Phenomenology. Phys Chem Chem Phys (PCCP) 1:3667–3671 Bousquet P (1957) Étude théorique des propriétés optiques des couches minces transparentes. Ann Phys Sér 13, 2:5–15 Cacace MG, Landau EM, Ramsden JJ (1997) The Hofmeister series: salt and solvent effects on interfacial phenomena. Q Rev Biophys 30:241–278 Csúcs G, Ramsden JJ (1998a) Interaction of phospholipid vesicles with smooth metal oxide surfaces. Biophys Biochim Acta 1369:61–70 Csúcs G, Ramsden JJ (1998b) Generalized ballistic deposition of small buoyant particles. J chem Phys 109:779–781 de Feijter JA, Benjamins J, Veer FA (1978) Ellipsometry as a tool to study the adsorption behaviour of polymers at the air-water interface. Biopolymers 17:1759–1772 Evans JW (1993) Random and cooperative sequential adsorption. Rev mod Phys 65:1281– 1329 Fernández A, Ramsden JJ (2001) On adsorption-induced denaturation of folded proteins. J Biol Phys Chem 1:81–84 Funatsu T, Harada Y, Tokunaga M, Saito K, Yanagida T (1995) Imaging of single fluorescent molecules and individual ATP turnovers by single myosin molecules in aqueous solution. Nature (Lond.) 374:555–559 Ghatak A, Thyagarajan K (1989) Optical Electronics. Cambridge University Press Guemouri L, Ogier J, Ramsden JJ (1998) Optical properties of protein monolayers during assembly. J chem Phys 109:3265–3268 Guemouri L, Ogier J, Zekhnini Z, Ramsden JJ (2000) The architecture of fibronectin at surfaces. J chem Phys 113; 8183–8186 Horváth R, Pedersen HC (2002) Demonstration of reverse symmetry waveguide sensing in aqueous solutions. Appl Phys Lett 81:2166-2168 Ishijima A, Kojima H, Funatsu T, Tokunaga M, Higuchi H, Tanaka H, Yanagida T (1998) Simultaneous observation of individual ATPase and mechanical events by a single myosin molecule during interaction with actin. Cell 92:161–171 Kurrat R, Ramsden JJ, Prenosil JE (1994) Kinetic model for serum albumin adsorption: experimental verification. J chem Soc Faraday Trans 90:587–590 Kurrat R, Textor M, Ramsden JJ, Boni P, Spencer ND (1997) Instrumental improvements in optical waveguide lightmode spectroscopy for the study of biomolecule adsorption. Rev Sci Instrum 68:2172–2176 Kurrat R, Wälivaara B, Marti A, Textor M, Tengvall P, Ramsden JJ, Spencer ND (1998) Plasma protein adsorption on titanium. Colloids Surfaces B 11:187–201 Levich VG (1962) Physicochemical Hydrodynamics. Prentice Hall, Englewood Cliffs, New Jersey


49

Lukosz W (1991) Principles and sensitivities of integrated optical and surface plasmon sensors for direct affinity sensing and immunosensing. Biosensors Bioelectronics 6:215–225 Mann EK (2001) Evaluating optical techniques for determining film structure:optical invariants for anisotropic dielectric thin films. Langmuir 17:5872–5881 Newton I (1730) Opticks, 4th edn. William Innys, London, Book 3, Qu. 29 Nishizaka T, Miyata H, Yoshikawa H, Ishiwata S, Kinosita K (1995) Unbinding force of a single motor molecule of muscle measured using optical tweezers. Nature (Lond.) 377:251–254 Mann EK, Heinrich L, Schaaf P (1997) Validation of the uniform thin-film approximation for the optical analysis of particulate films. Langmuir 13:4906–4909 Perutz M (1993) Protein Structure. W. Freeman, Oxford Ramsden JJ (1992) Observation of anomalous diffusion of proteins near surfaces. J phys Chem 96:3388–3391 Ramsden JJ (1993a) Review of new experimental methods for investigating random sequential adsorption. J statist Phys 73:853–877 Ramsden JJ (1993b) Calcium-dependence of laminin binding to phospholipid membranes. Biopolymers 33:475–477 Ramsden JJ (1993c) Concentration scaling of protein deposition kinetics. Phys Rev Lett 71:295–298 Ramsden JJ (1994) Experimental methods for investigating protein adsorption kinetics at surfaces. Q Rev Biophys 27:41–105 Ramsden JJ (1997) Optical biosensors. J molec Recog 10:109–120 Ramsden JJ (1998) Kinetics of protein adsorption. In: Malmsten M (ed) Biopolymers at Interfaces. Dekker, New York, Ch. 10 (pp 321–361) Ramsden JJ (1999) Molecular orientation in lipid bilayers. Phil Mag B 79:381–386 Ramsden JJ, Bachmanova GI, Archakov AI (1994) Kinetic evidence for protein clustering at a surface. Phys Rev E 50:5072–5076 Ramsden JJ, Li S-Y, Heinzle E and Prenosil JE (1995) An optical method for the measurement of number and shape of attached cells in real time. Cytometry 19:97–102 Ramsden JJ, Lvov YuA, Decher G. (1995) Optical and X-ray structural monitoring of molecular films assembled via alternate polyion adsorption. Thin solid Films 254:246–251; ibid. (erratum) 261:343–344 Rembe C, Dräbenstedt A, Heimes F (2005) An accurate new 3D-motion analyser for MEMS and bioMEMS. J Biol Phys Chem 5:37–40 Schaaf P, Talbot J. (1989) Surface exclusion effects in adsorption processes. J chem Phys 91:4401–4409 Spielman LA, Friedlander SK (1974) Role of the electrical double layer in particle deposition by convective diffusion. J Colloid Interface Sci 46:22–31 Talbot J (1996) Time dependent desorption: a memory function approach. Adsorption 2:89–94 Tiefenthaler K, Lukosz W (1989) Sensitivity of grating couplers as integrated-optical chemical sensors. J. opt. Soc. Am. B 6:209–220 Van Tassel PR, Guemouri L, Ramsden JJ, Tarjus G, Viot P, Talbot J (1998) A particle-level model of irreversible protein adsorption with a postadsorption transition. J Colloid Interface Sci 207:317–323 Viot P, Tarjus G, Ricci SM, Talbot J (1992) Random sequential adsorption of anisotropic particles. I. Jamming limit and asymptotic behavior. J chem Phys 97:5212–5218 Wiggins PM (2002) Enzyme reactions and two state water. J Biol Phys Chem 2:25–37 Yanagida T, Harada Y, Ishijima A. (1993) Nanomanipulation of actomyosin molecular motors in vitro: a new working principle. TIBS 18:319–323

3

Initial Adsorption Kinetics in a Rectangular Thin Channel, and Coverage-Dependent Structural Transition Observed by Streaming Potential Philippe Déjardin, Elena N. Vasina

Abstract. One fundamental parameter contributing to the analysis of protein–solid surface interactions is the adsorption kinetic constant ka . For a rectangular channel we give a global representation of the initial experimental adsorption constant k at distance x from the entrance channel, as a function of wall shear rate γ , through the variable 1.86 k(x/γ )1/ 3 . It is possible to visualize on a single graph the adsorption kinetic constant, the diffusion coefficient, depletion magnitude at the interface, and its thickness relative to that of the transport-limited Lévêque limit. With radiolabeled molecules (with a γ emitter like 125 I) and well-defined geometries, calibration does not require a known solute diffusion coefficient and is obtained from the abrupt variation of radioactivity at the buffer–protein solution change and vice-versa. Experimental data obtained for the system α-chymotrypsin/mica show that when surface coverage reaches some critical level, the streaming potential becomes almost independent of further interfacial concentration increase; this suggests an interfacial structural transition induced by interactions between adsorbed molecules. Several models based on protein–protein dipolar interactions are proposed.

3.1 Introduction Protein adsorption at solid–liquid interfaces (Andrade 1985; Brash and Horbett 1995) is important in many fields such as hemocompatible materials (Leonard et al. 1987), diagnostic kits (Malmsten et al. 1996), enzymatic activity (Quiquampoix and Ratcliffe 1992; Servagent-Noinville et al. 2000) and environmental hazards in mineral soils (Vasina et al. 2005). Natural and artificial vessels adsorb proteins from blood. Conformational changes or reactions at interfaces can induce series of biochemical reactions. In general, this type of phenomenon must be avoided as coagulation and complement Philippe Déjardin, Elena N. Vasina: European Membrane Institute, UMR 5635 (ENSCMUMII-CNRS), Université Montpellier 2, CC 047, 34095 Montpellier Cedex 5, France, E-mail: [email protected], [email protected] Principles and Practice Proteins at Solid–Liquid Interfaces Philippe Déjardin (Ed.) © Springer-Verlag Berlin Heidelberg 2006

52

P. Déjardin, E.N. Vasina

systems can be activated. Application to biosensors has been proposed (Mar et al. 1999). Surface hydrophilization via polymer pretreatment can inhibit or limit these phenomena in solid-phase diagnostics (Malmsten et al. 1996), in hemodialysis hollow fibers (Yan et al. 1992) or on polymer surfaces (Lee et al. 1990). Such polymer pretreatment is generally based on copolymers that contain poly (ethylene oxide) chains (Holmberg et al. 1997; Lee et al. 1989; Nitschke et al. 2000; Price et al. 2001; Tirelli et al. 2002; Wu et al. 2000; Xu and Marchant 2000) or phosphorylcholine moeities (Huang et al. 2005; Ishihara et al. 1999a, b; Iwasaki et al. 2003; Kojima et al. 1991; Nakabayashi and Williams 2003; Nakabayashi and Iwasaki 2004; Ueda et al. 1991; Ye et al. 2005; see Chaps. 10–12). In addition, protein adsorption modifies the interfacial charge density, or the electrokinetic potential at the interface, which can be deduced from streaming potential measurements. This technique was used also to study adsorption kinetics (Ethève and Déjardin 2002; Norde and Rouwendal 1990; Zembala and Déjardin 1994). One fundamental parameter contributing to the analysis of protein–solid surface interactions is the adsorption kinetic constant ka , which is related to the energy barrier the protein molecule has to overcome during adsorption mechanism. Determination of the adsorption constant, however, is not straightforward: The initial adsorption process can be controlled by transport or interfacial reaction as the two extreme limits; in addition, any intermediate case can exist where both the transport and the interfacial reaction have to be taken into account and their interplay accurately described. Under well-controlled laminar flow conditions, for instance, in a channel (Fig. 1) or tube with a radius much larger than the diffusion layer thickness, the experimental adsorption kinetic constant k(x) is compared to the Lévêque constant kLev (x) corresponding to a fully transport-controlled process in a rectangular channel, where x is the distance from the channel entrance to the observation point. Such comparison gives a qualitative estimation of the role of transport in the overall adsorption process. The rate of adsorption in the presence of low-concentration solutions, when the steady-state of the concentration profile C(x, y) has been established, can be written as:

∂ Γ(x, t) ∂ C(x, y) =D ∂t ∂y

= k(x) Cb = ka C(x, 0)

(1)

y=0

where Γ is the interfacial concentration, t is the time, D is the diffusion coefficient, Cb is the bulk solution concentration; ka is the adsorption kinetic constant at the interface, C(x, 0) is the solution concentration at distance x from the channel entrance and at y = 0, and k(x) is the kinetic constant of the overall process at x.

3 Adsorption kinetics and Coverage-dependent transition

53

Fig. 1. Channel with rectangular section. Flow occurs in the x direction. The distance to the wall is given by y, channel height by b, and fluid velocity profile by v(y) = γ y(1 − y/ b) , where γ is the wall shear rate

In case of the fully transport-controlled process (C(x, 0) = 0), the kinetic constant k at distance x from the entrance of the channel depends only on diffusion through the solute diffusion coefficient D and convection through the wall shear rate γ . According to Lévêque (1928) its expression is kLev (x) = 0.538 (D2 γ/ x)1/ 3 , a relationship that was also derived later (Levich 1962). When adsorption is controlled only by the interfacial reaction k ≈ ka and practically does not depend on x. Figure 2 illustrates the interfacial region of depletion in solution obtained by numerical simulations. For a given solute diffusion coefficient, the higher the adsorption kinetic constant, ka , the larger the interfacial depletion, and the larger the distance from the channel entrance, the larger the thickness, δ, of the depletion layer and the depletion magnitude at the surface. Hence the crossover length Lco (Eq. 5c) was introduced in the complete treatment (Déjardin et al. 1994), contrary to the simpler case of the rotating disk where δ is constant (Coltrin and Mitchell 2003; Levich 1962). We define the depletion d(x) at the interface: d(x) = 1 −

C(x, 0) Cb

(2)

As the steady-state adsorption rate is related to the slope of C(x, y) at the wall (Eq. 1): Cb − C(x, 0) ∂ Γ(x, t) =D = k(x) Cb = ka C(x, 0) ∂t δ(x)

(3)

54


Fig. 2. Stationary concentration profiles normalized to the bulk concentration Cb , for D = 6.0×10−7 cm2 s−1 , and wall shear rate 1,000 s−1 . Channel height is 100 µm. Left Threedimensional graph C(x, y) from the entrance to distance x = 2 cm. Right Concentration profile at x = 2 cm, with the tangent to the profile at the wall (dashed line, left scale), and parabolic velocity profile with the tangent at the wall (dash-dotted line, right axis). a ka = 1.0×10−4 cm s−1 ; b ka = 5.0×10−4 cm s−1 ; c ka = 1.0×10−2 cm s−1


55

we obtain: k(x) = ka [1 − d(x)] = D

d(x) . δ(x)

(4)

The Lévêque limit where depletion is complete corresponds to d → 1, kLev (x) = D/δLev (x) and δLev (x) = 1.859(Dx/γ )1/ 3 . When approaching this limit, k(x) 1), especially in studies concerning porous media (Adler and Thovert 1998). There is a direct connection between the ratio ka / kLev and this number. The difference originates from the introduction of the diffusion coefficient. Other expressions of Da are used that also take it into account (Bizzi et al. 2002; Coltrin and Mitchell 2003) for gas phases: Da = ka / (D/δ), where δ is the diffusion layer thickness. In the present problem, we can define a Damköhler number Da(x) = ka / kLev (x), with kLev (x) = D/δLev (x). As summarized earlier (Docoslis et al. 1999), the experimental determination of the adsorption kinetic constant ka has three major sources of difficulties: (1) mass transport, easier to take into account with the simple geometries of rectangular channels or circular tubes, (2) steric hindrance at the interface, which can be limited by using low-concentration solutions, and (3) determination of low interfacial concentration, which requires very sensitive techniques, usually using radioactive or fluorescently labeled molecules like in the total internal reflection fluorescence (TIRF) technique (Britt et al. 1998; Buijs et al. 1998; Kelly and Santore 1995; Malmsten et al. 1996; Rebar and Santore 1996; Robeson and Tilton 1996; Wertz and Santore 2002). Sophisticated optical methods such as surface plasmon resonance (SPR; Mar et al. 1999) and optical waveguide lightmode spectroscopy (OWLS; Calonder and Van Tassel 2001; Hook et al. 2002; Ramsden et al. 1995; Chaps. 1–2) provide the means to measure low interfacial concentrations without labeled molecules. The following data treatment can be useful for the SPR, reflectometry, ellipsometry, and OWLS techniques if the experimental flow cells have rectangular channel geometries. For instance, in OWLS experiments, a chamber tightened with an O-ring joint is not adapted to the present formulation as the channel width varies continuously and strongly near the inlet, which leads to a nonconstant wall shear rate. In this case the Lévêque limit formula is no longer applicable. The cell has to be adapted (Chap. 1) to achieve a constant width. Moreover, the height of the channel should be large enough to provide a linear velocity profile in the depletion layer, which is assumed in the Lévêque derivation, and its extension to finite ka (Déjardin et al. 1994).

56


For example, in Fig. 2 the channel height of 100 µm is too small to fulfill the criterion when D = 6.0×10−7 cm2 s−1 and γ = 1,000 s−1 . Another technique is the impinging jet on a flat substrate, which was also used in combination with TIRF to study the deposition of latex particles (Göransson and Trägardh 2000). Here we do not consider this kind of geometry. In the first part of the present work we recall derivation of the simple accurate approximations, providing an easy way to treat quantitative data; thus, one can deduce the protein diffusion coefficient D and protein interfacial adsorption kinetic constant ka from experimental initial constant k as a function of the variable (k/ kLev )D2/ 3 , where kLev / D2/ 3 does not depend on D. In the second part, we present the variation of the interfacial potential as a function of surface coverage and suggest the possible importance of protein–protein dipolar attractive interactions to create domains with an antiferroelectric order at high coverage.

3.2 The Initial Adsorption Constant and its Limit Expressions 3.2.1 The Local Initial Adsorption Constant k(x), its Limit Expressions and Approximation Full Solution and Limit Expressions Let us recall that the general expression (Dejardin et al. 1994) for the constant k(x) at distance x from the channel entrance as a function of ka and kLev is not easily accessible for the determination of ka . k(x) = ka g(X) ,

(5a)

with X= where

x Lco

Γ (2/ 3) Lco = 3 Γ (1/ 3)

(5b)

3

D2 γ k3a

(5c)

and g(X) = e−X + G(2/ 3, X) − G(1/ 3, X)

(5d)


with 1 G(n, x) = e−X Γ (n)

57

X

zn−1 ez dz

(5e)

0

Γ (n) is the usual gamma function (Abramovitz and Stegun 1972). The prime is added to avoid confusion with Γ used for the interfacial concen-

tration. It is preferable to have ka as a function of experimental k(x). It has been shown that rather good approximations of the general solution can be obtained at the two limits of the controls by transport and the interfacial reaction (Dejardin et al. 1994). These involve the inverse of the kinetic constants and, in fact, lead directly to the interpretation of the total resistance (time) of the adsorption process as being the sum of one resistance due to the transport and the other resistance due to the interfacial reaction. Close to the conditions of the transport-controlled process, −1 k−1 = k−1 Lev + 0.684ka ,

(6a)

while close to the conditions of the control by interfacial reaction −1 k−1 = 0.827k−1 Lev + ka

(6b)

Both expressions are similar to the simplest approximation, which considers no coupling between transport and the interfacial reaction: −1 k−1 = k−1 a + kLev

(7)

Equation 7 can be also obtained by assuming that the thickness of the diffusion layer (when C(x, y = 0) = 0, Lévêque model) is unaltered whatever the finite value of ka , say whatever the nonzero steady-state value of the volume concentration near the interface. Equation 7 corresponds to the work (Bowen and Epstein 1979) and matches Eq. 2.45 of Bowen et al. (1976). Approximation of k(x)/ ka as a function of k(x)/ kLev (x) If the linear approximations given in Eq. 6a and b are valid at to the two limits can be useful in practice, as it is easy to deduce ka from k(x) and kLev (x), they do not describe the entire domain as does the general expression (which writes k as a function of ka , not the reverse). Recently we have proposed (Déjardin and Vasina 2004) to approximate the general expression in Eq. 5a by the function y = f (u), where y = k/ ka and u = k/ kLev : f (u) =

(u − 1)(au − 1) (bu + 1)

(8)

with a = 0.451707 and b = −0.624713, to satisfy the two limits of Eq. 6a,b. The greatest relative variation between Eq. 8 and the complete calculation

58


Fig. 3. Adsorption of α-chymotrypsin (∇, dashed line, 10−2 M Tris; pH 8.6) onto mica. The normalization of the scales to the intercepts ka and D2/ 3 illustrates the magnitude of the depletion at the interface (Cy=0 / Cb ; right axis), where y is the distance to the interface, and the departure from the fully transport-limited process (k/ kLev ; top axis). For measurements performed at the same distance x, the experimental points corresponding to one given wall shear rate are positioned on a straight line passing through the origin. An example is provided with the points corresponding to experiments in 10−2 M (∇), 0.2 M ( ) and 0.5 M ( dash-dotted line) Tris, pH 8.6

is 1% around u = 0.8. Let us note that k(x)/ ka = C(x, 0)/ Cb , so the ordinate illustrates the depletion in solution at the interface, while u = k/ kLev estimates contribution of the transport. Figure 3 shows an example of these two items. Equation 8 can be transformed into a two-parameter fit (D and ka ) to experimental k(x, γ ) as a function of u = k(x/γ )1/ 3 / 0.538 = uD2/ 3 , which does not require the knowledge of D. k=

ka (u − D2/ 3 )(au − D2/ 3 ) D2/ 3 bu + D2/ 3

(9)

In such a representation, the intercept of the fit with the ordinate axis gives ka , while the intercept with the abscissa axis gives D2/ 3 (see Fig. 3). Figure 4 illustrates the magnitude of error in the adsorption kinetic constant ka and diffusion coefficient D that can occur when using Eq. 7. Equation 7 corresponds there to the straight line k = ka (1 − u / D2/ 3 ).


59

Fig. 4. Illustration of the maximum (Max.) error on adsorption constant ka and diffusion coefficient D when neglecting any coupling between transport and interfacial reaction (straight dash-dotted line; Eq. 7 in text), compared to exact solution (full curved line). The dashed lines demonstrate linear approximations at the two extreme limits

3.2.2 The Mean Adsorption Constant, its Limit Expressions and Approximation Mean Over the Full Length of the Channel In some cases it is necessary to find the average adsorption constant over the full sample length where laminar convection occurs. When the process is fully transport-controlled, the integration of the local Lévêque equation over the channel length L leads to = 0.808(D2 γ/ L)1/ 3 . When the contribution of the interfacial reaction is taken into account, the mean kinetic constant over the channel length L is given by (Valette et al. 1999): =

ka

Λ

Λ g(X)dX

(10a)

0

where:

3 L 3 ka Λ= = Lco 2Γ (2/ 3)

(10b)

Hence:

3Λ1/ 3 3Λ2/ 3 − − g(Λ) = 1+ Λ 2Γ (2/ 3) Γ (1/ 3) ka

(10c)

60


Close to the conditions of transport-controlled process, we have: −1 ≈ −1 + 0.913k−1 a

(10d)

while close to the conditions of control by the interfacial reaction −1 −1 ≈ k−1 a + 0.930

(10e)

with the total resistance being the sum of the two resistances (one due to the transport and the other one due to the interfacial reaction). As the numerical coefficients in both linear approximations are close to 1, the simple approximation −1 −1 ≈ k−1 a +

(11)

is rather good for the whole range of ka . The same route as that for the local adsorption constant led to the following approximation of /ka as a function of U =/< kLev >: (U − 1)(AU − 1) (12) F(U) = (BU + 1) where A = 0.203127 and B = −0.272759, to satisfy Eqs. 10d–e. The greatest relative variation between Eq. 12 and the complete calculation is 0.03% around U = 0.8. It can be used in the form = ka F(U) for a twoparameter fit (D and ka ) to the experimental value of as a function of U =(L/γ )1/ 3 / 0.808 = UD2/ 3 . k =

ka (U − D2/ 3 )(AU − D2/ 3 ) D2/ 3 BU + D2/ 3

(13)

Mean [k] Over a Restricted Length In experiments, the adsorption kinetics is always integrated over some length ∆x of the channel, between x − ∆x/ 2 and x + ∆x/ 2. We shall estimate the influence of ∆x on the numerical coefficients a and b in Eqs. 8 or 9. Referring to Eq. 8, the variable u = k (x)/ kLev (x) becomes u∗ = [k]/ [kLev ], where the star superscript and the brackets mean that the average is taken between x1 = x − ∆x/ 2 and x2 = x + ∆x/ 2. [k] is the actual measured average kinetic constant, which was assumed to be k(x) in Eq. 8, and [kLev ] = ∆x−1 (x2 2 − x1 1 ). i is the mean value of the Lévêque constant between x = 0 and x = xi . We adopt the same procedure looking for two limit expressions for a transport-limited process and interfacial reaction control (see Appendix). For small values of ε = ∆x/ x we obtain:

a ≈ a0 1 + 0.078 ε2 ; b ≈ b0 1 + 0.044 ε2 (14) where a0 and b0 are the values of the numerical coefficients determined in Eq. 8, which corresponds to ∆x → 0.


61

3.2.3 Experimental Results and Discussion A flow cell of adequate dimensions 4×11×1 cm (to be inserted in a detector of gamma radioactivity) consisted of two polymethylmethacrylate plates, between which were pressed two mica plates separated by a spacer. A top view of the cell is provided in Fig. 5. The data acquisition was performed using the same kind of procedure as for capillary geometry (Boumaza et al. 1992; Le and Déjardin 1998). For a detection window of width w and a channel height h, the radioactivity jump at the arrival of radiolabeled protein solution is Av ∼ whCb , while the surface radioactivity on the two faces (mica plates) is As ∼ 2wΓ. Therefore Γ(t) = (h/ 2)Cb As (t)/ Av and k(t) = (h/ 2)(dAs / dt)/ Av . Recently we studied (Vasina and Déjardin 2004) the adsorption of αchymotrypsin onto mica at different concentrations of Tris buffer at pH 8.6. Examples of the adsorption kinetics data are given in Fig. 6. The initial adsorption rate is linear with a protein solution concentration in the range considered (Fig. 7). In Fig. 3 some of our results emphasizing the peculiarities of the representation given by Eq. 9 are plotted. The scanning of adsorption with

Fig. 5. (A) Top view of the experimental cell showing the mica sheets constituting the two faces of the channel (scale in centimeters). Measurement of adsorption concerns only the central part of the cell, using appropriate lead shields (B). The cell is inserted in a γ radioactivity detector

62


Fig. 6. Experimental data of adsorption kinetics of α-chymotrypsin on mica recorded for wall shear rate 120 s−1 and different bulk α-chymotrypsin concentrations, Cb : 10 µg ml−1 (•); 5 µg ml−1 (); 2.5 µg ml−1 ()

Fig. 7. Initial adsorption rate of α-chymotrypsin on mica as a function of solution concentration Cb at pH 8.6, Tris buffer (concentration 10−2 M). Mean slope (dashed line) k ≈ 1.2 × 10−4 cm s−1

convection is viewed as straight lines passing through the origin with their slope increasing with wall shear rate γ as 0.538 (γ/ x)1/ 3 . The application of Eq. 9 to the adsorption of α-chymotrypsin onto mica in the rectangular channel provides two parameters: the diffusion coefficient D and the adsorption kinetic constant ka (for instance, in 10−2 M Tris buffer, pH 8.6: D = 8.8 ± 0.7×10−7 cm2 s−1 and ka = 4.4 ± 0.5×10−4 cm s−1 ). The twoparameter fit was performed using SigmaPlot®.


63

The depletion (Eq. 2) with its complement to unity, say k(x, γ )/ ka or C(x, 0, γ )/ Cb , is directly visualized in Fig. 3. Moreover, given the curvature of the function k(u ), the thickness of the depletion layer is always smaller than that of the Lévêque model, as k(x, γ ) = Dd(x, γ )/δ(x, γ ) with d(x, γ ) = 1 in the Lévêque limit, therefore δ(x, γ )/δLev (x, γ ) = d(x, γ )/ [k(x, γ )/ kLev (x, γ )]. Both terms of the ratio are noted in Fig. 3. In our experiments, the detection occurs between 2.5 and 5.5 cm from the channel entrance, therefore ε = 0.75. Taking into account this range of integration (Eqs. 9 and 14), we obtain for Tris buffer 10−2 M at pH 8.6, corrections of –1% for the protein diffusion coefficient D and +6% for its adsorption kinetic constant ka . When comparing the present treatment to other works in the literature, we should mention that we do not assume any value for the protein diffusion coefficient. The calibration is provided by the increase in radioactivity when the radiolabeled protein solution arrives in a channel of known geometry or by its drop when the radiolabeled solution is replaced by buffer. We do not assume any Lévêque regime with a known diffusion coefficient, as is sometimes assumed in fluorescence (TIRF) experiments (Wertz and Santore 1999, 2002).

3.3 The Structural Transition with Increasing Interfacial Concentration As can be seen in Fig. 6, the adsorption rate slightly increases at about half the maximum interfacial concentration. Such an increase could be the signature of a structural change at the interface. It was recognized for many years that protein crowding at the interface leads to profound changes there because of the increasing probability of the protein–protein interactions (Andrade 1985). In the present case, such a change in the adsorption rate could also be caused by the conditions of flow in the channel, associated with the early saturation of the surface upward of the point of examination. Simulations suggest, however, that this hypothesis should be rejected, as it would require an adsorption constant much higher than that deduced from the previous analysis (Vasina and Dejardin 2004). To describe the conformational changes (see Chap. 6) and/or orientation, many models can be proposed, among them the side-on/end-on process, the reverse process, and the spreading of the (soft) protein on the surface. Techniques like neutron reflectivity (Su et al. 1998), ellipsometry (Bae et al. 2005; Cuypers et al. 1978; McClellan and Franses 2005; Poksinski and Arwin 2004; Seitz et al. 2005; Werner et al. 1999), and scanning angle reflectometry (Ladam et al. 2002; Schaaf and Déjardin 1988) are able to provide information about

64


the orientation of an ellipsoid-modeled molecule. Circular dichroism and differential scanning calorimetry are also useful in the study of the protein conformations (Giacomelli and Norde 2001; Norde and Zoungrana 1998; Vermeer and Norde 2000; Vermeer et al. 2001; Voegel et al. 1987; Zoungrana et al. 1997). Infrared attenuated total reflection (Noinville et al. 2002a,b, 2003) provides data about possible changes in the tertiary structure (see Chap. 6). The TIRF technique was refined to take into account the dependence on the emission properties of the fluorophore with its ionization state or with the electric potential (Daly et al. 2003; Robeson and Tilton 1996). If the ionic strength is chosen in such a way that the interfacial electric potential varies over a distance comparable to the protein size, then the signal will be dependent upon the orientation of the molecule. Indeed the interfacial transition is easily observed and is concomitant with the change in kinetic regime.

3.3.1 Observation by Streaming Potential Measurement of streaming current (Daly et al. 2003) and streaming potential (Ethève and Dejardin 2002; Vasina and Dejardin 2004) during adsorption can be useful to obtain information on the orientation of molecules at interfaces. The variation of streaming potential with protein interfacial concentration exhibits a change at some critical concentration for the α-chymotrypsin/mica system at pH 8.6 in 10−2 M Tris (Fig. 8). The variation in zeta potential was deduced from the streaming potential, Es , as a function of time at a defined pressure differential, ∆P, via the classical Smoluchowski relationship (Hunter 1981), with ∆Es = Es − Es0 , where Es0 is the asymmetry potential,

∆Es ηλ0 (15) ∆P ε0 εr where λ0 is the electrolyte solution conductivity, η is the solution viscosity, ε0 is the vacuum dielectric permittivity and εr is the relative permittivity. ς=

The quantitative analysis of the data, however, is not simple as initially, when transport is important, the interfacial concentration is not uniform. Indeed, the Smoluchowski relationship assumes a uniform charge density along the channel walls. For another system – lysozyme/silica capillary; pH 7.4, 10−2 M phosphate buffer (Ethève and Dejardin 2002) – we observed the same kind of variation and assumed that the streaming potential resulted from the mean charge density along the capillary and estimated by numerical simulations the average interfacial concentration. This did not


65

Fig. 8. Variation of the ζ potential, relative to the initial ζ0 potential (–95 mV) of the bare mica, as a function of α-chymotrypsin interfacial concentration. Solution concentration Cb = 2.5 µg ml−1 in Tris 10−2 M, pH 8.6. Wall shear rate 120 s−1 . The full line represents the expected variation, based on the behavior at small θ, for particles bearing only one type of charge

lead to any dramatic changes in the observation of the transition. After the transition, local and average concentrations become much closer. When comparing to similar works with random deposition of uniformly charged particles (Zembala and Adamczyk 2000) where such transition was not observed, it was concluded that the transition could be the result of the peculiar nature of the proteins. Contrary to the particles, proteins bear the two types of charge, positive and negative. Moreover, the charge distribution on the protein surface is not uniform. Let us analyze the present system along the same lines as the ones used for particles. As a first approximation, α-chymotrypsin can be assumed globally neutral because its isoelectric point is 8.2. For particles, the variation of ζ/ζ0 with coverage was described as obeying a sum of an exponential function and a linear one (Zembala and Adamczyk 2000).

ςp ς = e(Ci0 ϑ) + Cp0 ϑ (16a) ς0 ς0 where ζ0 and ζp are the potentials of the bare surface and uniformly charged

particle, respectively, in the presence of buffer. After additional control experiments, another expression was proposed (Zembala 2004): ςp ς = e(Ci0 ϑ) + 1 − e(−Cp0 ϑ) (16b)

ς0

ς0

66


When the particle radius r is much larger than the Debye length κ−1 , then Ci0 = −10.21 and Cp0 = 6.51. The dependence of those coefficients on rκ has been established (Zembala et al. 2001). As the dimensions of α-chymotrypsin are 5.1 × 4 × 4 nm3 , with r = 2−2.5 nm and κ−1 = 6.3 nm, we have rκ = 0.3 − 0.4, thus Ci0 ≈ −4.5−5. As the net protein charge is almost zero, according to the model, the variation of ζ/ζ0 should follow an exponential decay, exp(−Ci0 θ), corresponding to the flow attenuation or substrate charge screening by the deposited neutral protein particles. We found the constant 1.8 instead of 4.5 (Fig. 8). The passage from local to mean coverage by simulations leads to the slightly lower value of 1.7. The smaller decrease in ζ/ζ0 with coverage, compared to what should be expected from the model, may have two explanations: the distribution of charges on the protein surface and/or the nonrandom adsorption (Dabros and van de Ven 1993). It is clear, however, that whatever the mechanism, proteins and homogeneously charged particles have different qualitative behaviors at high coverage. The reason is probably the protein–protein interaction associated with the mobility of proteins on the surface. As α-chymotrypsin possesses a high dipole moment (483 D; Antosiewicz and Porschke 1989), the reasonable model of parallel dipoles at low ionic strength and low coverage, based on the ionic interaction between the mica and the protein, would lead at high coverage to a strong repulsion component between parallel dipoles, when the distance between like charges become shorter. As experiments show that the maximum interfacial concentration for proteins, and especially here for α-chymotrypsin, is generally near the full close-packed monolayer, some specific mechanism should occur.

3.3.2 Different Models Figure 9 illustrates the two models of rearrangements of the dipoles roughly normal or parallel to the surface to obtain an attractive component between adsorbed proteins. We proposed the model of the ferroelectricantiferroelectric arrangement normal to the surface to explain the change in variation of the ζ potential (Vasina and Dejardin 2004; Fig. 9, left). Saving almost the same orientation of the proteins with respect to the surface (Robeson and Tilton 1996), cooperative clustering could appear also by dipolar interactions in the direction parallel to the surface (Fig. 9, right). The complete analysis would, however, require a careful description of the balance between attractive and repulsive electrostatic components, probably in connection with the local concentration of small ions. Such ferroelectric to antiferroelectric order transition was already considered for


67

Fig. 9. Schematic representation of possible ordered domains at the interface, with a top view of the charge repartitions. a Alternate orientation of the dipoles normal to the surface for intermolecular attraction, but with a repulsive component in front of the surface. b Alternate orientation of the dipoles parallel to the surface for intermolecular attraction, but with a repulsive intermolecular component between the likely oriented dipoles normal to the surface. Ellipsoid or cylinder representation is arbitrary

small dipolar molecules like carbon monoxide and halogenated methane (Burns and Dennison 1998; Burns et al. 2004). Recently, it was shown by TIRF and streaming-current measurements on the lysozyme/silica system at high wall shear rates that the TIRF overshoot did not correspond to any particular event in the interfacial potential (Daly et al. 2003). Therefore, it appears that the concomitance of the changes in kinetic regime and interfacial potential is not always the rule. One important phenomenon behind the process is surface diffusion, which has to be taken into account. Different kinds of behavior are probably linked to the relative rate of arrival of molecules compared to their surface diffusion rate (see Chap. 2).

3.4 Conclusion We proposed simple relationships in the channel geometry to describe the passage from a fully transport-controlled process to an interface-controlled one when increasing convection (say wall shear rate γ ) for a study at some distance x from the entrance, as well as for the average value over the full or partial length of the channel (or wall). The experimental data of the initial experimental kinetic constant k(x, γ ) are plotted as a function of

68


(k/ kLev )D2/ 3 , where kLev is the Lévêque limit and D the diffusion coefficient. The adsorption kinetic constant ka and the protein diffusion coefficient to power 2/3 can be easily determined as the intercepts of the fit curve with the ordinate and abscissa axes, respectively. The normalizations of ordinates and abscissae to the intercepts provide directly the magnitude of the depletion in the liquid phase at the interface, and the relative evaluation to the transport-limited process. An example is given for the system α-chymotrypsin/mica at pH 8.6. The variation of ζ potential as a function of surface coverage exhibits a change of regime, this suggests a new arrangement of adsorbed protein molecules above a threshold of surface occupation. The possibility of dipolar attractive interactions between the proteins is considered to create domains with an antiferroelectric order at high coverage. Acknowledgements. We are grateful to R. Souard and P. Montels (EMI) for manufacturing the flow cell, and to D. Cot (EMI) for the photo. This work was supported by “Programme Ecologie Quantitative” of the French Ministry of Research, and performed within the framework of collaboration between Centre National de la Recherché Scientifique (France) and Kazan State University (Russia), project 12889.

Appendix We study the influence of the length of examination ∆x around the average mean distance to the channel entrance x, through the parameter ε = ∆x/ x, on the numerical parameters a and b determined in Eq. 8. The two local limit expressions, valid near the control by the interfacial reaction (Eq. 6b) −1 or by transport (Eq. 6a), can be written as k−1 (x) = ca k−1 a + cL kLev (x), where ca and cL are numerical coefficients (Dejardin et al. 1994; Dejardin and Vasina 2004). In what follows, the average between x1 = x − ∆x/ 2 and x2 = x + ∆x/ 2, is indicated by the use of brackets []. 1. k ≈ ka > 10−2 M), a direct spectroscopic measurement was used. The optical density of the solution was recorded at 240 nm and the resulting concentration was calculated using a value of 39.4/mol/cm (Aebi 1984) for the molar extinction coefficient of H2 O2 . Poly(pyrrole-biotin) Film Characterization The structure of the film was assessed by scanning electron microscopy. The polymerization of the pyrrole–biotin monomer was carried out on a carbon felt. Figure 25 shows photographs of polymer deposition on the carbon surface. The latter illustrates that the electrochemical addressing of polymer films allows the spatially controlled functionalization of surfaces whatever their shape and size are. We recently demonstrated by gravimetric measurements that the immobilization of avidin on biotinylated polypyrrole films via bioaffinity interactions provides a compact avidin monolayer (Cosnier et al. 2001). Figure 25C shows this specific binding event of avidin onto polymerized biotin leading to a smoothed polypyrrole–air interface. Peroxidase Immobilization Amount of Immobilized Enzyme

The amount of POD immobilized on the carbon felt after electropolymerization were determined by enzymatic activity measurements. The

9 Biocatalytic textile microreactors

231

Fig. 25. Micrographs of carbon felt before (A) and after modification by electropolymerization of pyrrole-biotin monomer (B) and deposited on platinum surface (C) before (left) and after (right) complexation with avidin

enzyme solutions were analyzed before and after contact with the carbon tissue and the amount of enzyme actually grafted in the polymer matrix was calculated in the same way as described previously in the experimental section. The results demonstrate that 1.6 mg of enzyme was anchored in the polymeric matrix, corresponding to 83% of the initial amount of enzyme (2 mg) put in contact with the carbon felt modified by the polymer. The amount of fixed active enzyme in the membrane is around 4.1 mg/cm3 , which corresponds to approximately 0.1 mg/cm2 of the effective surface. Taking into account the fact that the theoretical maximum surface coverage for a close-packed avidin monolayer corresponds to 3.3−5×1012 molecules/cm2 (Cosnier 1999), the maximum amount of an-

232

C. Innocent, P. Seta

chored POD should be 0.22−0.33 mg/cm2 . The specific anchoring of POD by avidin–biotin bridges thus constitutes an efficient approach for the functionalization of the whole structure of a carbon felt. The specific recognition between avidin and biotinylated enzyme allows a strong increase in enzyme grafting as compared to other matrices (Cosnier and Innocent 1993). Thus, the complexation between biotin and avidin appears to be limited by a steric interaction, whereby the smaller the enzyme size, the more efficient the immobilization. The same result has been reported for enzyme immobilization in the case of polypyrrole derivatives for biosensor elaboration (Cosnier and Innocent 1993). Membrane Working

An equimolecular aqueous solution of H2 O2 and pyrogallol (5 × 10−3 M) passed through the POD membrane and was analyzed by absorption spectroscopy. In the course of the transfer across the membrane, H2 O2 is reduced by the immobilized POD, which in turn oxidizes pyrogallol into purpurogallin (see Eqs. 1 and 2). The amount of purpurogallin detected corresponds to the quantity of H2 O2 consumed by the enzymatic reaction. The chemical yield of the membrane process corresponds to the ratio between the number of purpurogallin molecules produced and the number of H2 O2 molecules introduced into the column. A value of 0.8 was obtained for this ratio. This value, although satisfying by itself, can be improved by a recirculation of the feed solution through the enzymatic membrane. After a recirculation time of 30 min (which corresponds to six passages of the feed solution volume through the membrane), all of the initial H2 O2 was oxidized by the enzymatic catalytic membrane. Figure 26 shows the rate of consumption of H2 O2 (initial concentration 5 × 10−3 M) versus time, whereby the exit solution is continuously injected into the column, thus allowing a constant flow through the enzymatic membrane. The remarkable efficiency of this H2 O2 degradation process is explained by the peculiar molecular interaction process used for the enzyme attachment in the immobilization procedure. In this way the enzyme was neither altered nor denatured as a result of avidin–biotin complexation and the catalytic activity was maintained in the membrane. The measurement of the specific activity of the free POD in solution related to pyrogallol transformation into purpurogallin indicates 200 U/mg of enzyme (a value not very far from that notified by the supplier, 240 U/mg). The membrane enzyme activity determined in the same experimental conditions was 117 U/mg. This value corresponds to 58% of free enzyme activity. Thus, the remaining activity of the immobilized enzyme is markedly higher than those reported previously with other immobilization methods based on electrogenerated polymer films (only 4% of free peroxidase activity; Coche-Guérente et al.


233

Fig. 26. Conversion ratio versus time of H2 O2 solution (10−3 M) with peroxidase membrane (membrane thickness 1 cm, flow rate 1 ml/min)

1995) and 9% of free polyphenol oxidase activity (Cosnier and Innocent 1993) is maintained in the polymer matrix. This clearly illustrates the advantages offered by a soft attachment of the enzyme molecules, which preserves their conformational flexibility and hence their activity. Dependence on pH

The optimum pH range for reactivity of the catalytic membrane was investigated with H2 O2 (5 × 10−3 M) as a substrate and a circulation time of 15 min (flow rate: 2 ml/min). In the pH range of 4.7–8.3 the enzyme membrane shows a typical bell-shaped response; t. The maximum efficiency is obtained for pH 6.8 (Fig. 27). This value is in good agreement with that determined with free POD in solution (Worthington 1988). The kinetic experiments presented below were performed at pH 6.8, which corresponds to the pH of maximum efficiency for the immobilized enzymes. Kinetic Parameters Kinetic analysis was carried out by recording the solution substrate concentration of the at the column exit. According to the investigation reported previously (Wang et al. 1996), the immobilized enzyme reactivity in the carbon felt was estimated by considering the residence time of the substrate solution in the membrane. The mean residence time, tr , of solute in the porous material can be calculated using the following equation: tr = L/ J

(8)

234


Fig. 27. Effect of pH on the enzyme activity of the membrane

where L is membrane thickness and J is volume flux of the solution; J is equal to the ratio of flow to membrane surface. The reaction rate is calculated as V = ∆C/ tr , where ∆C is the concentration decrease of H2 O2 in the flowing solution. Peroxidase Membrane Case

Figure 28A depicts the relationship between the reaction rate and substrate concentration in the case of a POD membrane. The rate follows a classical enzymatic kinetic curve. The first step corresponds to a linear part (weak substrate concentration). For high substrate concentrations, saturation of the reaction rate is observed and the maximal value obtained is due to saturation of active enzyme sites. Figure 28B shows the corresponding Lineweaver-Burk plot for the immobilized POD membrane. From the obtained slope (KM / VM ) and intercept (1/ VM ) for the linear relationship, both KM and VM are estimated to be 1.45 × 10−4 M and 1.8 × 10−3 M/s, respectively. Catalase Membrane Case

In order to use the enzyme membrane for H2 O2 dismutation, immobilization of catalase has been achieved. A simplified scheme for the mechanism underlying the catalase reaction may be described by the following two equations (Chance 1948; Nicoholls and Schonbaum 1963): cat(OH)4 + H2 O2 ↔ cat(OH)3 OOH + H2 O cat(OH)3 OOH + H2 O2 → cat(OH)4 + O2 + H2 O

(9) (10)


235

Fig. 28. Reaction rate of H2 O2 dismutation with POD-polypyrrole membrane (A) and Lineweaver-Burk plot of the POD-polypyrrole membrane (B); residence time, tr = 42 s, membrane thickness 1 cm, flow rate 1 ml/min

where cat(OH)4 is native catalase and cat(OH)3 OOH the oxidized enzyme complex. A kinetic study was carried out both with the peroxidase and the catalase membrane in order to understand the effect of immobilization by avidin-biotin technology. In some cases, the treatment of the kinetics re-

236


Table 6. Kinetic parameters: Michaelis constants (VM and KM ) for catalase and peroxidase k*[E] Free catalase Biotinylated catalase Immobilized catalase Immobilized peroxidase

1×10−2 9×10−4 5×10−2

k (/s/mg)

KM (mol/l)

0.125 0.011 0.031

3.26×10−2 9.5×10−3 7 1.45×10−4

VM (mol/l/s) 0.11 3×10−2 1.8×10−3

sults with a classical Lineweaver-Burk plot gives a negative value for the extrapolated reaction rate. This suggests that the maximum reaction rate of enzymatic reaction is negative, which does not make any sense. For instance, for cases in which zero or negative y-intercepts are obtained, the KM is much larger than the total experimentally accessible substrate concentration and the simplified Michaelis-Menten model does not apply. The same phenomenon has been observed previously with another enzymes such as lipase (Freeman et al. 2000). This is the reason why in Table 6 we preferred to use the catalytic constant (k) to compare the activity of the different enzymes. The catalytic constant (expressed as per seconds per milligram normalized to a constant amount of enzyme) for the immobilized catalase is defined as the ratio of the slope of the curve reaction rate versus substrate concentration. This value is four times lower than that of the free catalase. This result is in good agreement with those values obtained for catalase immobilized on cellulose (Eremin et al. 1995). However, it is interesting to note that for catalase, biotinylation of the enzyme decreases the rate by a factor of about 10, whilst immobilization enhances the rate as compared to that of the biotinylated enzyme in solution. The KM value drastically increases when catalase is immobilized by avidin-biotin recognition. This value is not a significant physical parameter as it corresponds to an unrealistic concentration of H2 O2 . At these high concentrations the catalase should be totally deactivated by the substrate (Vasudevan and Weiland 1990). Enzyme immobilization, for instance in polymeric hydrogels (Arica et al. 1999), has been shown to exhibit a large KM value. A strong increase in the value of KM means that the kinetic parameters are affected by enzyme immobilization, presumably due to the steric effects, which significantly decrease the catalytic reaction rate. Membrane Stability The enzymatic membrane was stored at 4 ◦ C in buffer solution. Assays were performed at different storage times. Figure 29 illustrates membrane stability measured in the case of POD immobilization as the H2 O2 concentration converted to purpurogallin, starting from a 5 × 10−5 M equimolar H2 O2 -


237

Fig. 29. Enzyme activity (in %) of the POD-membrane (•) and catalase membrane ( ) after storage at 4 ◦ C in buffer solution. Activities were measured with hydrogen peroxide concentration of 10−3 M after circulation time of 20 min (flow rate: 1 ml min−1 )

pyrrogallol aqueous solution flowing across the membrane. In the case of catalase immobilization, the same operational stability with an initial H2 O2 concentration of 2 × 10−2 M, was recorded. No significant decrease in enzyme activity was observed after 140 days and 8 h of cumulative dismutation reaction with a flow rate of 2 ml/min. The enzymatic membrane was remarkably stable compared to membranes made by other immobilization methods (Coche-Guérente et al. 1995; Cosnier 1999; Cosnier and Innocent 1993) and thus it should be noted that the enzyme was actually irreversibly immobilized. In the case of catalase immobilization, the stability is also satisfying; a loss of enzyme activity was only observed after 140 days of storage in buffer solution at 4 ◦ C. Conclusion In this work the efficiency of the immobilization method based on avidinbiotin technology linked to an intermediate electropolymerized polymer layer onto carbon fibers for the building of a catalytic filtration membrane has been demonstrated. The enzyme activity of this new type of membrane along with the kinetic parameters have been determined in the case of the dismutation of H2 O2 . These rate parameters are not too much affected by the immobilization process, as the loss of reactivity as compared to free

238


enzymes was shown to be moderate. Moreover, the long-term stability of the catalytic activity of this type of reactive membrane makes it a promising candidate for practical application in membrane technology for liquid transformations.

9.3.2 Electrocatalysis Coupling with Enzyme-Conducting Textile Catalytic Reactivity The electrical conducting properties of the carbon felt grafted with redox enzymes with the aid of a polypyrrole polymer layer, allow the electrochemical coupling or electrochemical transformation of the reactants involved in the activity of redox enzymes linked to such immobilization supports. As the enzyme support acts as an electrically conducting membrane, coupling via an electrochemical process is a simple means to intervene in the catalytic enzyme reaction. The electrochemical reaction has been used successfully for the regeneration of a substrate of the enzyme (converting a product of the enzyme activity into substrate) or even for the production of substrate in the vicinity of the enzyme These two processes have been demonstrated with peroxidase. In a first example the cosubstrate of the peroxidase reaction, for instance Fe(CN)4− 6 , which is oxidized by peroxidase, is reduced electrochemically on the conducting textile biased to a negative electric potential. In the second example, H2 O2 , a substrate of peroxidase, is electrochemically produced by the reduction of oxygen, applying a reducing electrical potential to the conducting textile (in the latter case the textile acts as a working electrode for oxygen reduction). Electrochemical Regeneration of an Enzyme Cosubstrate: Peroxidase Activity A carbon felt modified by electro-deposition of a polypyrrole biotin layer to which biotinylated horseradish peroxidase (HRP) is attached by the molecular recognition process involving the biotin-avidin interaction (described in Chap. 1) has been introduced in a simple reactor where a solution of the enzyme reactants (H2 O2 and Fe(CN)6 K4 ) flow across the enzyme-modified carbon felt (Fig. 30). The conducting enzyme support is connected to a potentiometer as a working electrode biased to a potential at which the reduction of Fe(CN)6 K3 occurs. The reaction is initiated by the addition of a low concentration of Fe(CN)6 K4 to start the enzyme catalytic reaction, and the enzyme turnover is maintained by the electrochemical coupling, as the cosubstrate of the HRP: Fe(CN)6 K4 is electrochemically regenerated from the oxidized form Fe(CN)6 K3 formed in the reactor. The electrochemical coupling process is shown in Fig. 30.


239

Fig. 30. Scheme of the working principle of peroxidase (HRP) immobilized on carbon felt covered with a polypyrrole biotin polymer. The HRP is attached to the polymer surface by biotin-avidin molecular recognition linkage. The regeneration of the redox mediator Fe(CN)4− 6 is achieved biasing the carbon felt at –0.2 V versus an Ag/AgCl electrode

The variation of the Fe(CN)6 K3 concentration is followed by absorption spectroscopy at 305 nm. Depending on the initial concentration of the Fe(CN)6 K4 cosubstrate, for a given initial H2 O2 concentration, the production of oxidized Fe(CN)3− 6 species increases (see Fig. 31), which is an indication of the activity of the immobilized peroxidase. After a given time, the concentration of Fe(CN)6 K3 decreases in the reactor as a result of the electrocatalytic coupling effect regenerating the cosubstrate. This effect is also evidenced by the reduction in the concentration of H2 O2 as the enzyme activity is started. The concentration of Fe(CN)6 K3 affects the initial slope of the ferricyanide production as well as that of the related H2 O2 dismutation, but the slope ratios do not vary exactly in the same proportion. The more probable explanation for this is the diffusion-limited process of access of reactants to the immobilized enzymes sites in the core of the grafted textile. However, the data shown in Fig. 32 indeed confirm the electrocatalysis coupling, as the amount of H2 O2 converted is much higher than that of ferricyanide initially introduced in the reactor. After about 50−60 min, as 1 mole of H2 O2 necessitates 1 mole of Fe(CN)6 K3 to be converted into products, the initial ferricyanide concentration introduced in the reactor has been approximately regenerated 100 times by the electrical reduction coupling at the carbon felt surface.

240


Fig. 31. Variation with time of the potassium ferricyanate absorbance for different initial concentrations of Fe(CN)6 K4 in 10−3 M H2 O2 solution (the carbon felt is electrically biased at –0.2 V versus Ag/AgCl electrode)

Fig. 32. Change in the H2 O2 substrate concentration during the HRP reactivity under coupling with electrocatalysis


241

Electrochemical Production of an Enzyme Substrate Another illustration of the interest of the electrical connection of the enzyme textile support is the possible production, for instance, of the substrate of a peroxidase activity. The electrochemical production of H2 O2 is used in water treatment processes (Drogui et al. 2001), and it can be used similarly here for the working function of HRP immobilized onto a carbon felt. The direct reduction of oxygen onto the carbon felt surface was achieved by biasing the substrate at –0.5 V/ECS, where the oxygen reduction occurs according to the classical electrochemical reaction: 1 O2 + 2H+ + 2e− → H2 O2 2

(11)

The scheme shown in Fig. 33 depicts the working conditions of the enzyme reactor used in association with the electrochemical production of the HRP substrate, the chosen cosubstrate being pyrogallol, and oxygen fed in the reactor by gas bubbling. As in the case of the regeneration of the HRP cosubstrate described in the previous paragraph, the coupling efficiency is evidenced by the data shown in Fig. 32. One of the products of the enzyme catalytic process, purpurogallin, is clearly formed from the pyrogallol fed into the reactor. The purpurogallin concentration depends on that of the initially introduced pyrogallol, showing that the H2 O2 substrate is electrochemically produced in sufficient

Fig. 33. Scheme of the catalytic reactions on the surface of a modified textile

242


amount from oxygen fed into the reactor, and does not limit the kinetics of the enzyme reaction. Conclusion In conclusion, the two examples mentioned in this section demonstrate that the immobilization of enzymes by a molecular recognition process onto textiles composed of electron-conducting fibers can be coupled to electrocatalysis. This coupling is efficient, even if in the described data some kinetic limitations are noticeable, due presumably to limited diffusion processes of species in the vicinity of the enzyme reactive sites in the textile core. However, as these experiments have been achieved in a very simple laboratory enzyme reactor, better optimization of the reactor design should emphasized this point. Acknowledgements. We are very indebted to N. Bouhaine at the Department of Chemistry, Badji Mokhtar University, BP 12, 23000 Annaba, Algérie, and to A. Gherrou at the Faculty of Chemistry, USTHB, 16111 Bab Ezzouar, Algérie for their contributions to the experiments involving trypsin and pepsin grafted onto cotton. The authors also thank Professor Zhi-Kang XU and Dr. Zhen-Mei LIU at the Institute of Polymer Science, Zhejiang University, Hangzhou, 310027 People’s Republic of China, for their collaboration in the domain of enzyme modifications of membranes in the frame of the Advanced Research Program (PRA 03-05) between France and China.

References Aebi HE (1984) Methods of Enzymatic Analysis, vol. 3, (3rd edn). Verlag Chemie, Berlin, p 273 Albayrak N, Yang ST (2002) Immobilisation of Aspergillus oryzae galactosidase on tosylated cotton cloth. Enzyme Microb Technol 31:371–383 Arica MY, Oktem H A, Oktem Z, Tuncel SA (1999) Immobilization of catalase in polyisopropylacrylamide-co-hydroxyethylmethacrylate thermally reversible hydrogels. Polym Intern 48:879–887 Buschle-Diller G, El Mogahzy Y, Inglesby MK, Zeronian SH (1998) Effects of scouring with enzyme, organic solvents, and caustic soda on the properties of hydrogen peroxide bleached cotton yarn. Text Res J 68:920–929 Chance B (1948) The enzyme-substrate compounds of catalase and peroxides Nature 161:914–917 Coche-Guérente L, Cosnier S, Innocent C, Mailley P (1995) Development of amperometric biosensors based on the immobilization of enzymes in polymer films electrogenerated from a series of amphiphilic pyrrole derivatives. Anal Chim Acta 311:23–30 Coche-Guerente L, Deronzier A, Mailley P, Moutet JC (1994) Electrochemical immobilization of glucose oxidase in poly(amphiphilic pyrrole) films and its application to the preparation of an amperometric glucose sensor. Anal Chim Acta 289:143–153


243

Coche L, Moutet JC (1987) Catalysis of 1,2-dibromo-1,2-diphenylethane reduction on platinum and carbon felt electrodes coated by polypyrrole films containing 4,4 bipyridinium groups. J Electroanal Chem 224:111–124 Confort AR, Albert EC, Langer R (1989) Immobilized enzymes cellulose hollow fibers: part I: immobilization of heparinase. Biotechnol Bioeng 34:1366–1373 Confort AR, Mullon CJP, Langer R (1988) The influence of bond chemistry on immobilized enzyme systems for ex vivo use. Biotechnol Bioeng 32:554–563 Cosnier S (1999) Biomolecule immobilization on electrode surfaces by entrapment or attachment to electrochemically polymerized films. A review. Biosens Bioelectron 14:443–460 Cosnier S, Innocent C (1993) A new strategy for the construction of a tyrosinase-based amperometric phenol and o-diphenol sensor. Bioelectrochem Bioenerg 31:147–160 Cosnier S, Perrot H, Wessel R (2001) Biotinylated polypyrrole modified quartz crystal microbalance for the fast and reagentless determination of avidin concentration. Electroanalysis 13:971–982 Cosnier S, Stoytcheva M, Senillou A, Perrot H, Furriel RPM, Leone FA (1999) A biotinylated conducting polypyrrole for the spatially controlled construction of amperometric biosensor. Anal Chem 71:3692–3697 Dalmon J-A (1997) Catalytic membrane reactors. In: Ertl G, Knözinger H, Weitkamp J (eds) Handbook of Heterogeneous Catalysis. Wiley-VCH, Weinheim, pp 1387–1398 Dean RB (1991) Processes for water reclamation. Waste Manage Res 9:425–430 Dejean E, Laktionov E, Sandeaux J, Sandeaux R, Pourcelly G, Gavach C (1997) Electrodeionization with ion-exchange textile for the production of high resistivity water: influence of the nature of the textile. Desalination 114:165–173 Deng HT, Xu ZK, Wu J, Ye P, Liu ZM, Seta P (2004) A comparative study on lipase immobilized polypropylene microfiltration membranes modified by sugar-containing polymer and polypeptide. J Mol Catal B Enzym 28:95–100 Devi S, Guthrie JT, Beddows CG (1990) The immobilization of trypsin onto irradiated oxidized natural rubbers converted to the natural rubber-g-Co-HEMA system. Radiat Phys Chem 36:703–707 Drogui P, Elmaleh S, Rumeau M, Bernard C, Rambaud A (2001) Hydrogen peroxide production by water electrolysis: application to disinfection. J Appl Electrochem 31:877–882 Eltsefon BS, Vengerova NA, Vysotina TA, Rudman AR, Yermakova LN, Irkley VM, Ryabchenko AS, Kuznetsova NA, Lykovykh LM (1988–89) Structure and transport properties of the Ultracell membrane for hemodialysis. Biomater Artif Cells Artif Organs 16:967–975 Eremin AN, Otyutskii SV, Metelissa DI (1995) Properties of catalase immobilized on cellulose in aqueous and micellar media. Kinet Catal 36:776–784 Freitas dos Santos LM, Pavasant P, Strachan LF, Pistikopoulos EN, Livingston AG (1997) Membrane attached biofilms for waste treatment – fundamentals and applications. Pure Appl Chem 69:2459–2469 Freeman KS, Tang TT, Shah RDE, Kiserov DJ, Gown LB (2000) Activity and stability of lipase in AOT reversed micelles with bile salt cosurfactant. J Phys Chem B 104:9312–9316 Giacomini C, Villarino A, Franco-Fraguas L, Batista-Viera F (1998) Immobilization of galactosidase from Kluveromyces lactis on silica and agarose: comparison of different methods. J Mol Catal B Enzym 4:313–327 Gonzalez-Garcia J, Bonete P, Exposito E, Montiel V, Aldaz A, Torregrosa-Marcia R (1999) Characterization of a carbon felt electrode: structural and physical properties. J Mater Chem 9:419–426 Green MJ, Hill HAAO (1984) Chemistry of dioxygen. Methods Enzymol 105:3–22 Huang TC, Chen DH (1992) A study of removal of urea from aqueous solution with immobilized urease and electrodialysis. J Chem Tech Biotechnol 55:191–199

244


Huang TC, Chen DH (1993) Coupling of urea hydrolysis ammonium removal in an electrodialyzer with immobilized urease. Chem Eng Commun 1205:191–201 Institut Textile de France (1987) Celluloses greffees Capt’ion: un echangeur d’ions à cinetique rapide, document technique Iqbal SS, Bruno MW, Bronk JG, Batt BV Chambers CA (2000) A review of molecular recognition technologies for detection of biological threat agents. Biosens Bioelectron 15:549– 578 Ishihara K, Shinozuka T, Hanazaki Y, Iwasaki Y, Nakabayashi N (1999) Improvement of blood compatibility on cellulose hemodialysis membrane: IV. Phospholipid polymer bonded to the membrane surface. J Biomater Sci Polym Ed 10:271–282 Kaewprasit C, Hequet E, Abidi N, Gourlot JP (1998) Application of methylene blue adsorption to cotton fiber specific surface area measurement: Part I. Methodology. J Cotton Sci 2:164–173 Kamath N, Melo JS, D’Souza SF (1988) Urease immobilized on polyethyleneimine cotton cloth. Appl Biochem Biotechnol 19:251–258 Kumar SD, Kulkarni AV, Kalyanraman R, Krishnamoorthy TS (1997) Whole blood glucose determination using glucose oxidase immobilized on cotton cheese cloth. Anal Chim Acta 338:135–140 Lee S, Lueptow R M (2001) Membrane rejection of nitrogen compound. Environ Sci Technol 35:3008–3018 Marchand Brynaert J (1999) Surface functionalization of polymer membranes. In: Smith Sorensen T (ed) Surface Chemistry and Electrochemistry of Membranes, Surfactant Science Series, vol 79. Marcel Dekker, New York, pp 91–124 Nicoholls P, Schonbaum GR (1963) Catalases. In: Boyer P, Lardy H, Myrback K (eds) The Enzymes, vol. VIII (2nd edn). Academic Press, New York, pp. 147–225 Nolan Etters J (1998) Emerging opportunities for enzyme use in textiles. Colourage Annu 87–92 Rao S, Anderson KW, Bachas LG (1999) Controlled layer-by-layer immobilization of horseradish peroxidase. Biotechnol Bioeng 65:389–396 Schussel LJ, Atwater JE (1995) A urease bioreactor for water reclamation aboard manned spacecraft. Chemosphere 30:985–994 Schmidt B (1996) Membranes in artificial organs. Artif Organs 20:375–380 Vasudevan PT, Weiland RH (1990) Deactivation of catalase by hydrogen peroxide. Biotechnol Bioeng 36:783–789 Vo-Dinh, Cullum BM (2000) Biosensors and bioships: advances in biological and medical diagnostics. Fresenius J Anal Chem 336:540–551 Wang H Y, Kobayashi T, Saitoh H, Fujii N (1996) Porous polydimethyl-siloxane membranes for enzyme immobilization. J Appl Polym Sci 60:2339–2346 Worthington BC (1988) Peroxidase. In: Worthington Enzyme Manual. Worthington Biochemical Co. Freehold, 254 Yakup Arica M (2000) Epoxy-derived pHEMA membranes for use bioactive macromolecules immobilization: covalently bound urease in a continuous model system. J Appl Polym Sci 77:2000–2008 Yamazaki H, Cheok RKH, Fraser ADE (1984) Immobilization of invertase on polyethyleneimine-coated cotton cloth. Biotechnol Lett 6:165–170

10

Approaches to Protein Resistance on the Polyacrylonitrile-based Membrane Surface: an Overview Ling-Shu Wan, Zhi-Kang Xu, Xiao-Jun Huang

Abstract. Protein adsorption and/or deposition at the surface of polymeric membranes play important roles in membrane separation processes. Although polyacrylonitrile-based membranes have been used successfully in many fields, surface modifications to improve the protein resistance and hemocompatibility have received considerable interest. Various methods such as copolymerization, grafting, physical adsorption, biomacromolecule immobilization, and biomimetic modification, have been explored to build a friendly microenvironment for proteins, especially enzymes, at the membrane surface. Herein, the behaviors of proteins at the surfaces of polyacrylonitrile-based membranes are reviewed.

10.1 Introduction Polyacrylonitrile (PAN) exhibits good thermal stability, mechanical strength, and membrane formation properties, and microporous PAN membranes are commercially available. However, its relatively poor hydrophilicity and biocompatibility hinder this type of membrane in further applications. Just like other ultrafiltration membranes, PAN-based membranes face an unavoidable problem with regard to their applications: a progressive decline in flux and a change of membrane selectivity. This phenomenon, more commonly known as membrane fouling, is attributed mainly to concentration polarization and protein fouling. Protein fouling, including adsorption/deposition on the membrane surface and pore blockage, generally results in a largely irreversible decline of flux. Many researchers have therefore provided important insights into the effects of interactions between proteins and membranes or between proteins. On the other hand, PAN-based membranes have also been used as bloodcontacting materials in hemodialysis, plasmapheresis, plasma fractionation, leukofiltration, and artificial livers (Krasteva et al. 2002; Sun et al. Ling-Shu Wan, Zhi-Kang Xu, Xiao-Jun Huang: Institute of Polymer Science, Zhejiang University, Hangzhou 310027, People’s Republic of China, E-mail: [email protected] Principles and Practice Proteins at Solid–Liquid Interfaces Philippe Déjardin (Ed.) © Springer-Verlag Berlin Heidelberg 2006

246

L.-S. Wan et al.

2003). During these processes protein adsorption occurs rapidly, within seconds to minutes after the first blood contact. In addition to the aforementioned disadvantages of flux decline and change in selectivity, protein adsorption can trigger the cascading chemical reactions of clotting factors, which is followed by platelet adhesion and activation of the coagulation pathways, leading to thrombus formation (Dee et al. 2002). It seems that proteins at the membrane surface stand at the center of membrane filtration processes, especially for the blood-contacting membranes. Proteins, not only model proteins of membrane fouling such as bovine serum albumin (BSA), but also a special protein type, enzymes, must be taken into account. Protein molecules should not be thought of as rigid structures. As we know, proteins are flexible chains that have been coiled, folded, and bent to assume a particular three-dimensional conformation. Changes in the microenvironment of proteins can alter the conformation of the molecules. Likewise, proteins experience structural alterations during interactions with solid surfaces, such as charged or noncharged membrane surfaces. As for protein fouling, protein denaturation and aggregation may result in radically different deposition compared to native proteins. Thus, keeping protein adsorption at a low level, thus allowing only little changes in its conformation, should be beneficial to membrane fouling. Many approaches have been explored to build a friendly and biocompatible interface between protein and membrane. Among them, surface modifications to increase the hydrophilicity and to introduce steric hindrance and/or a biomimetic layer on the membranes have received considerable attention. Herein, we will review some recent progress in the field of protein resistance at PAN-based membrane surfaces, as effected by various surface modification methods.

10.2 Copolymerization Procedures Acrylonitrile can be easily copolymerized with a variety of comonomers (Krasteva et al. 2002). Copolymerizations of hydrophilic/functional monomers such as maleic acid (Nie et al. 2004a,b), N-vinyl-2-pyrrolidone (Groth et al. 2002; Krasteva et al. 2002; Wan et al. 2005), α-allyl glucoside (Xu et al. 2003, 2004), 2-hydroxyethyl methacrylate (HEMA; Huang et al. 2005a,b), acrylamides (Musale and Kulkarni 1996, 1997), and N-vinylimidazol (Godjevargova et al. 2000) with acrylonitrile have been performed to improve the properties of PAN-based membranes. Modifications through copolymerization are capable of improving the bulk property of polyacrylonitrile to some extent. Furthermore, the surface properties of the membrane can also be greatly improved, because the amount of incorporated comonomers

10 Proteins at Polyacrylonitrile-based Membrane Surface

247

can be accurately tuned, and the hydrophilic moieties of the comonomer may migrate to the membrane surface to minimize the surface energy when used in an aqueous environment. Carbohydrates exist in many forms and play important roles in natural living systems. Their highly hydrophilic characteristics together with their innate compatibility with biomolecules may meet the demands of protein resistance and have led to considerable interest in the synthesis of their polymers. We have synthesized the copolymer of acrylonitrile with α-allyl glucoside using the process of water-phase precipitation copolymerization (Xu et al. 2003). Copolymer membranes with different contents of α-allyl glucoside were prepared using a phase inversion process (Xu et al. 2004). The formula of α-allyl glucoside and the water contact angle of the membranes are shown in Figs. 1 and 2, respectively. BSA was chosen as a model protein to investigate the protein resistance of these copolymer membranes and typical results are shown in Fig. 3. It is well known that the hydrophobic interaction between the membrane surfaces and proteins plays a very important role in the nonselective adsorption of protein onto those membrane surfaces. Materials that possess a hydrophilic surface usually show relatively low nonselective adsorption for proteins or cells. Carbohydratecontaining polymers are highly hydrophilic materials, a property that can be confirmed by water-contact angle measurements; however, some recognize the biomolecules or cells because of the “cluster effect”. Therefore, the hydrophilicity and recognition function of the carbohydrate moieties have different effects on the adsorption of proteins or cells. It can be seen from Fig. 1 that the adsorbed amount of BSA decreases almost linearly with increases in α-allyl glucoside content in the copolymer: the higher BSA concentration may lead to the larger amount of BSA adsorbed. The decrease in BSA adsorption can be ascribed mainly to the improvement in hydrophilicity by the carbohydrate moieties for the membrane surface. Macrophage adhesion was also performed on these membranes to evaluate their biocompatibility; the results are shown in

Fig. 1. Chemical structure of α-allyl glucoside

248

L.-S. Wan et al.

Fig. 2. Relationship between the contact angle and the content of α-allyl glucoside (AG) in the acrylonitrile/α-allyl glucoside (AN/AG) copolymer membranes

Fig. 3. Adsorption of bovine serum albumin (BSA) onto AN/AG copolymer membranes


249

Fig. 4. Relative macrophage number versus AG content in AN/AG copolymer membranes

Fig. 4. It was clearly demonstrated that the number of macrophages that adhered onto the copolymer membranes decreased sharply when compared with that on the polyacrylonitrile membranes. However, it seems that the macrophage adhesion number increased again when the content of α-allyl glucoside exceeded around 20 wt %. This might be due to the “cluster effect” of the carbohydrate moieties in the copolymer chains. It was reported that the carbohydrate density strongly affects the specific interactions between carbohydrate moieties and cells, and that the binding affinity is drastically enhanced by multivalent carbohydrate ligands (the “cluster effect”). Poly(N-vinyl-2-pyrrolidone) has excellent biocompatibility with living tissues. N-vinyl-2-pyrrolidone is also a potential comonomer for the chemical modification of polyacrylonitrile membranes by copolymerization. Groth and coworkers (Groth et al. 2002; Krasteva et al. 2002) synthesized copolymers of acrylonitrile with N-vinyl-2-pyrrolidone using a solution copolymerization process. They used this type of copolymer membrane as culture substrate for human skin fibroblasts for the development of an artificial skin. The attachment, morphology, and growth of hepatocytes on this copolymer membrane were also investigated. They proposed that there was no simple relationship between the wettability of the membrane and its ability to support cell adhesion and function. It was also suggested that this copolymer containing N-vinyl-2-pyrrolidone can be considered as an interesting substrate if sufficient numbers of hepatocytes are seeded to promote sufficient functionality for a biohybrid liver support system.

250

L.-S. Wan et al.

Fig. 5. Adsorption of BSA at different concentrations (2 g/l open triangles; 5 g/l filled squares) on acrylonitrile/N-vinyl-2-pyrrolidone copolymer films with various NVP content

Acrylonitrile/N-vinyl-2-pyrrolidone copolymers with different contents of N-vinyl-2-pyrrolidone were also synthesized by us using a water-phase precipitation copolymerization process. It was found that the introduction of N-vinyl-2-pyrrolidone into PAN did not change the water contact angles much, and this is in accord with the findings of Groth et al. (2002). However, as shown in Fig. 5, BSA adsorption was remarkably depressed with increasing content of N-vinyl-2-pyrrolidone in the copolymer. The results of platelet adhesion also confirmed the improvement in the biocompatibility of PAN afforded by the incorporation of N-vinyl-2-pyrrolidone. One can envisage that the protein fouling on this copolymer membrane surface might be reduced or even partly eliminated. How do polymers containing N-vinyl-2-pyrrolidone achieve excellent biocompatibility? Hayama et al. (2004) insisted that the biocompatibility of membranes containing N-vinyl-2-pyrrolidone was dependent on both the number of N-vinyl-2-pyrrolidone units and their surface structures. They proposed that the higher the regularity of the polymer particle structure in the wet condition, the lower the wet:dry ratio surface roughness and the larger the wet:dry ratio of the polymer particle diameter. That is, the more the polymer particles swell as a result of wetting, the greater the biocompatibility. We have synthesized the copolymers of acrylonitrile with maleic acid using the water phase precipitation copolymerization process (Nie et al. 2004c), which were denoted as PANCMA. Ultrafiltration hollow-fiber membranes of these copolymers with different molecular weights and contents of maleic acid were also prepared by a dry–wet phase inversion process. The


251

carboxyl groups on the membrane surfaces were then converted into anhydride groups by refluxing the membranes in acetic anhydride; the resultant membrane was denoted as PANCMAn. Although the amount of protein adsorbed onto both PANCMA and PANCMAn membranes were large, it was obvious that the former, with a BSA adsorption of 2.31 g/m2 , was more protein resistant than the latter, with a BSA adsorption of 4.02 g/m2 . Based on the fact that the water contact angle on the PANCMA membrane was 55 ◦ C, which was smaller than that on the PANCMAn membrane (65◦C, this different adsorption behavior of BSA might be ascribed to differences in the surface hydrophilicity. Musale and Kulkarni (1996) investigated the fouling reduction in membranes fabricated from PAN and acrylonitrile-based copolymers with increasing acrylamide content. They suggested that membranes containing acrylamide were more hydrophilic, had a smaller dispersion force component of the surface energy and a smaller negative zeta potential than those prepared from PAN. The effect of the surface chemistry of these membranes with similar pore sizes was studied through the ultrafiltration of BSA as a function of feed pH. It was found that the PAN membrane exhibited relatively low BSA transmission rates throughout the pH range 4.0–7.5. However, the acrylamide-containing membranes exhibited marked increases in BSA transmission at pH values above the protein isoelectric point (pH 4.8). This might be ascribed to the combined effects of improved hydrophilicity/reduced dispersive surface energy and less electrostatic repulsion between the copolymer membranes and BSA. On the contrary, at pH values below the isoelectric point (pH 4.8), all of the membranes had a low BSA transmission due to strong adsorption resulting from attractive electrostatic interactions. They also found that the permeate flux and flux recovery increased for all membranes with increasing feed pH. This might be partly attributed the decreasing electrostatic attraction and increasing electrostatic repulsion between BSA and the membrane surface. At any given pH, the data comparison of permeate fluxes and flux recoveries indicated less fouling of the surface in the case of the acrylamide-containing membrane. In other words, the acrylamide-containing PAN membrane had a greater protein resistance. Hemoglobin, a protein with a similar molecular weight to BSA but that is more hydrophobic, was also used by Musale and Kulkarni (1997) as a model protein in the ultrafiltration of PAN membranes containing acrylamide. In this case, because the hydrophobic interactions were strong, the rejection was relatively constant and the flux exhibited a minimum at the isoelectric point, while it increased monotonically through the copolymer membranes. It was also proposed that hemoglobin would be adsorbed preferentially on the membrane surface during the ultrafiltration of mixed proteins of hemoglobin and BSA.

252

L.-S. Wan et al.

A copolymer of acrylonitrile and N-vinylimidazol was synthesized by Godjevargova et al. ( 2000) and its ultrafiltration membrane was used for glucose oxidase immobilization. They found that the enzyme bound on the copolymer membrane had high relative activity (91%). One can envisage that the membrane containing N-vinylimidazol provided the immobilized enzyme with a good microenvironment and could optimize the conformation of enzyme for catalysis.

10.3 Poly(ethylene glycol) Tethering Surface graft polymerization is one of the most commonly used methods of grafting, including radical, radiation, plasma, and ultraviolet initiated graft polymerization. In general, grafting modification changes the surface properties of a polymer membrane in a defined selective way while preserving its bulk/macroporous structure. The physicochemical properties of the membrane surfaces change to a great extent after modifications. Therefore, interactions with solutes such as proteins with a membrane can be changed accordingly, which may cause protein resistance. Many researchers have developed various technologies to realize surface grafting. Meanwhile, a great variety of monomers were chosen to improve the protein resistance of PAN membranes. Ulbricht and Belfort (1996) studied the surface and permeation properties of PAN and polysulfone membranes after treatment with low-temperature plasma and subsequent grafting with HEMA. They indicated that PAN ultrafiltration membranes hydrophilized by simple water/He plasma treatment exhibited a strongly reduced impact of protein adsorption on membrane fouling, and hence enabled higher filtrate fluxes with the same protein retention. Based on the results of BSA ultrafiltration, they also proposed that the reduction of protein surface adsorption by increased membrane surface hydrophilicity could diminish the tendency of protein fouling, thus compensating for a loss in hydraulic permeability due to the grafting modification. Being different from some monomers such as acrylamide and acrylic acid, poly(ethylene glycol) (PEG) is well-known for its extraordinary ability to resist protein adsorption because of its hydrophilicity, large excluded volume, and unique coordination with surrounding water molecules in an aqueous medium (Nie et al. 2003, 2004b, d; Xu et al. 2005). Surface-grafted PEG has also rendered ultrafiltration membranes resistance to protein fouling. In order to improve the protein resistance and blood compatibility of the PANCMA membrane, the carboxyl groups on the membrane surfaces were converted into anhydride groups by refluxing the membranes in acetic anhydride, and then these anhydride groups are subjected to an esterifica-


253

Fig. 6. Illustration of the preparation of acrylonitrile/maleic acid copolymer membrane and its grafting with poly(ethylene glycol) (PEG)

tion reaction with PEG containing hydroxyl end groups, as illustrated in Fig. 6. For surface grafting modification, the grafting degree or the amount of grafted polymer is important to the surface properties of membrane. The grafting degree can be tuned by changing the conditions of modification during radiation-, plasma-, or ultraviolet-initiated graft polymerization. In order to investigate the effect of the grafting degree of PEG on the protein resistance of the membrane, PANCMA membranes with different contents of maleic acid were fabricated and then PEG was immobilized onto these membrane surfaces according to the procedure described above. The typical characteristics of the PANCMA and PEG-grafted membranes are summarized in Table 1. It can be seen from Table 1 that the grafting degree of PEG increased with the content of maleic acid, and the hydrophilicity (depicted as water contact angle) increased in the same order. To investigate the filtration performances of this series of membranes, dynamic protein adsorption experiments were carried out. Typical results for the permeation fluxes of water and BSA solution through all membranes are shown in Figs. 7 and 8. The symbols, Jw0 and Jp are defined as the flux Table 1. Typical characteristics of the acrylonitrile/maleic acid copolymer (PANCMA) and poly(ethylene glycol) (PEG)-grafted membranes. CA Water contact angle, GD grafting degree Polymera

Contentb

CA (◦ Cc )

Polymerd

GDe

CA (◦ C)

PANCMA04 PANCMA07 PANCMA11

3.69 7.48 11.45

48.2 42.4 37.1

PANCMA04-g-PEG400 PANCMA07-g-PEG400 PANCMA11-g-PEG400

7.19 13.17 16.84

31.4 24.2 21.1

a,b The number following “PANCMA” refers to the mol content of maleic acid in the copoly-

mer c Measured by the sessile drop method d PEG400-grafted PANCMA membrane with different maleic acid contents e Calculated from the weights of a membrane before and after grafting reaction (wt%)

254

L.-S. Wan et al.

Fig. 7. Permeation fluxes of deionized water (Jw0 , hatched column) and BSA solution (Jp , shaded column) through: PANCMA04 (1), PANCMA07 (2), PANCMA11 (3), PANCMA04g-PEG400 (4), PANCMA07-g-PEG400 (5), and PANCMA11-g-PEG400 membranes (6). The number following “PANCMA” and “PEG” refers to the mole content of maleic acid and PEG, respectively, in or grafted onto the copolymer, respectively. PANCMA acrylonitrile/maleic acid copolymer

of deionized water and BSA solution, respectively. As can be seen from Fig. 7, for all PANCMA membranes, both Jw0 and Jp membranes increased with the content of carboxyl groups, and the PEG-grafted membranes had higher fluxes than all nongrafted membranes. However, it was interesting that the Jw0 and Jp of the PANCMA11-g-PEG400 (i. e., the membrane with the largest grafting degree) of PEG, were smaller than that of PANCMA07g-PEG400 (i. e., a membrane with a moderate grafting degree). This may be due to the pore plugging of the membrane with the high graft degree of PEG. Figure 8 shows the percent of total fouling [(1 − Jp / Jw0 ) × 100], the flux recovery following water cleaning [(Jw1 − Jp )/ (Jw0 − Jp ) × 100, where Jw1 denotes water permeation flux after water cleaning], and the flux recovery as a result of chemical cleaning [(Jw2 −Jw1 )/ (Jw0 −Jp )×100, where Jw2 denotes the water permeation flux after chemical cleaning]. It can be clearly seen that the PEG-grafted membranes possessed lower total fouling and higher flux recovery. Figure 9 shows the results of static protein adsorption for membranes with different contents of carboxyl groups or grafting degree of PEG. It


255

Fig. 8. Flux changes of: PANCMA04 (1), PANCMA07 (2), PANCMA11 (3), PANCMA04-gPEG400 (4), PANCMA07-g-PEG400 (5), and PANCMA11-g-PEG400 (6) membranes during filtration (hatched columns) and after water cleaning (black columns) and chemical cleaning (crosshatched columns)

can be seen that, the higher the content of carboxyl groups or PEG, the smaller the amount of static BSA adsorption. In addition, the PEG-grafted membranes had a greater degree of protein resistance than the PANCMA membranes. This result implied that the grafting modification with PEG rendered the membrane somewhat protein resistant. To investigate the protein resistance of the PEG-grafted membrane, PEGs with various molecular weights were grafted onto the PANCMA membrane. Water contact angle and water absorption measurements were used to characterize the relative hydrophilicity of the membrane surface. It can be seen from Fig. 10 that the PEG400-grafted membrane showed the greatest degree of water absorption and the smallest water contact angle. This is interesting and can be simply explained as follows. Although the PANCMA04g-PEG200 membrane has the largest number of PEG chains per unit area, namely grafting density, the chain is short and of relatively low hydrophilicity. On the other hand, it is more difficult to graft PEG600 and PEG1000 onto the membrane surface; therefore, the hydrophilicity of the corresponding membranes also decreases. It seems that to obtain an optimal protein-resistant membrane, PEG of a moderate molecular weight should

256

L.-S. Wan et al.

Fig. 9. Relationship between static BSA adsorption and maleic acid (MA) content in PANCMA for the original (squares) and PEG-grafted (triangles) membranes

Fig. 10. Effects of the molecular weight of PEG on the water contact angle (open circles, left y axis) and water absorption (filled triangles, right y axis) of: polyacrylonitrile (PAN; 1), PANCMA3 (2), PANCMA04-g-PEG200 (3), PANCMA04-g-PEG400 (4), PANCMA04-gPEG600 (5), and PANCMA04-g-PEG1000 (6) membranes


257

Fig. 11. Relationship between the molecular weight of PEG and static BSA adsorption on: PAN (1), PANCMA04 (2), PANCMA04-g-PEG200 (3), PANCMA04-g-PEG400 (4), PANCMA04-g-PEG600 (5), and PANCMA04-g-PEG1000 (6) membranes

be suitable because of the balance between the grafting density and the length of the hydrophilic chain segments. Figure 11 shows the results of static protein adsorption experiments on membranes grafted with different molecular weights of PEG. It can be seen that little protein is adsorbed on PEG-grafted membranes.

10.4 Physical Adsorption One of the commonest methods of modifying the surface properties of a membrane is to coat that membrane, which has the desired bulk properties, with an agent that has the desired surface properties. Mostly, coating is based on the physical adsorption of an agent on the membrane surface. Hence, the stability and durability of the coating agent is one of the key factors. Déjardin et al. (Etheve et al. 2003; Thomas et al. 2000;Valette et al. 1999; Yan et al. 1992) reported a series of studies on modification of the physical adsorption properties of PAN-based membranes. Two types of poly-

258

L.-S. Wan et al.

Fig. 12. Schematic representation of the structure of the copolymer and terpolymer derived from acrylonitrile, poly(ethyleneoxide) methacrylate and trimethylaminoethyl chloride methacrylate

mers, namely copolymers and terpolymers, were synthesized as depicted in Fig. 12. In the case of preadsorption with the terpolymer, platelet accumulation was considerably reduced. The effect on albumin and fibrinogen adsorption following preadsorption with the copolymer was also investigated. They found that whatever the pH of preadsorption, passivation efficiency was zero for copolymers with a high content of poly(ethyleneoxide) side chains, while a maximum reduction in adsorbance of 60% was attained with a copolymer containing 79 mol% acrylonitrile. They proposed that a higher content of acrylonitrile in the copolymer might improve the anchoring of the copolymer onto the membrane surface and therefore prevent protein adsorption more effectively than in the case of copolymers with a lower content of acrylonitrile, which is more readily displaced from the surface by proteins. Based on this consideration, the terpolymer was also synthesized and physically adsorbed onto the membrane. A reduction of at least 87% in fibrinogen adsorbance was obtained. They have also reported that although this water-soluble terpolymer might be leached out during these experiments, such a leaching should be low enough that it would not affect the prevention of protein adsorption. They also found that the levels of activation of blood components by the preadsorbed PAN-based membrane were lower than those of controls. Based on these results, it can be concluded that this physical adsorption modification did afford protein resistance. In the study of Thomas et al. (2000), positively-charged poly(ethyleneimine) was physically adsorbed onto various PAN-based hemodialysis membranes, probably creating a repelling and water-swollen layer for proteins. Participating in the contact-phase activation of the endogenous blood-coagulation cascade, high-molecular-weight kininogen was chosen as a model protein to study the protein resistance of these membranes. It was found that both kininogen adsorption and contact-phase activa-


259

tion was greatly reduced, irrespective of the pH value (between 7.0 and 7.8). They (Etheve et al. 2003) also performed lysozyme adsorption on a hemodialysis sulfonated PAN membrane with and without preadsorbed poly(ethyleneimine) on the external faces. They found that over a long time period, the total adsorbed amount of lysozyme adsorbed was unaffected by preadsorption of the membrane with poly(ethyleneimine). However, it was found that the poly(ethyleneimine) layer led to significantly slower kinetics of adsorption at pH 7.2 in Tris buffer. Thus, it can be inferred that surface preadsorption by a high-molecular-weight poly(ethyleneimine), while preventing the activation of contact-phase activation, did not inhibit the high capacity of the membrane to adsorb small proteins. Both cytochrome C and α-lactalbumin are low-molecular-weight proteins. The former is positively charged in an aqueous solution at pH 7.4 and the latter is negatively charged. Based on the results of the adsorption of these two low-molecular-weight proteins on different hemodialysis membranes, Valette et al. (1999) suggested that the observed differences in the adsorbance of proteins were mostly due to the different microstructure, chemical nature, and surface charge of the membrane.

10.5 Biomacromolecule Immobilization Since PAN is one of the most important polymeric materials used in bloodcontacting devices, efforts have been taken to build a blood-compatible surface. A well-known anticoagulant, heparin, which can catalytically increase the formation rate of antithrombin III and inhibit thrombin and some other coagulating proteases, is commonly used to treat patients who submit to hemodialysis. Heparin is a mixture of variably sulfated polysaccharide chains composed of repeating units of d-gluosamine and either l-iduronic or d-glucuronic acids, as well as some other biomacromolecules such as chitosan, insulin. Together with some proteins, heparin has been proved to be an effective agent in curtailing thrombosis and is effective when immobilized onto polymer surfaces. In order to modify a PAN membrane to endow antibacterial activity and anticoagulation activity, Lin et al. (2004) immobilized the surface of PAN membrane with a chitosan/heparin complex, denominated as PANC/H. The anticoagulation activity was evaluated using protein adsorption, platelet adhesion, and coagulation time including activated partial thromboplastin time, prothrombin time, fibrinogen time, and thrombin time. Their immobilization procedure is shown in Fig. 13. They reported that the hydrophilicity of the membrane after biomacromolecule-induced immobilization increased slightly. However, the be-

260

L.-S. Wan et al.

Fig. 13. Chemical scheme of biomacromolecule-immobilized PAN membranes. A Direct reaction of heparin, collagen, or albumin. B Chitosan (C)/heparin (H) conjugate reaction. EDC N,N-(3-dimethylaminopropyl)-N’-ethyl carbodiimide, GA glutaraldehyde, PAN-A activated PAN membrane, PAN-H heparin-grafted PAN membrane, PAN-PC collagen-grafted PAN membrane, PAN-PA albumin-grafted PAN membrane, PAN-C chitosan-grafted PAN membrane, PAN-C/H chitosan/heparin-complex-grafted PAN membrane

havior of protein adsorption onto membranes depended significantly on the surface characteristics, hydrophilicity, roughness, surface charge and chemistry. As shown in Table 2, the adsorbed amount of human serum albumin (HSA) and human plasma fibrinogen (HPF) on PAN-C/H was reduced to 38% and 26%, respectively, of those of PAN. The isoelectric point of HSA and HPF in the blood are 4.8 and 5.5, respectively; thus these proteins carry a negative charge in the normal blood circumstance (pH 7.4). Lin et al. (2004) proposed that, after hydrolysis, PAN-A carries a negative charge (-COO− ), thus the protein adsorption was reduced. On the other hand, chitosan is a weak base with a pKa of 6.5, thus PAN-C carries positive charge (-NH+3 ) at pH 7.4. The adsorption of these two proteins was promoted because of the electrical attraction; however, when PAN-C was heparinized, the repulsive SO−3 and COO− on heparin reduced the adsorp-


261

Table 2. Platelet adhesion and plasma protein adsorption on the PAN-based membranes surfaces Membrane type

Platelet adhesion numbers (cells/mm2 )

HAS adsorption (µg/cm2 )

HPF adsorption (µg/cm2 )

PAN PAN-A PAN-C PAN-H PAN-C/H PAN-PA PAN-PC

4810±38 3811±28 6689±57 1770±15 1623±16 2584±26 6050±40

279.7±14.9 261.7±12.5 288.3±13.9 154.4±11.9 145.4±12.5 – –

564.3±16.1 501.3±15.2 615.3±14.3 211.3±12.7 147.9±15.6 237±14 588±29

tion of proteins. To sum up, they considered that electrical attraction or repulsion might play a key role in the adsorption of protein. Lin et al. (2004) found that both the protein adsorption and platelet adhesion on the heparin-immobilized PAN membrane surface had the same trend; that is, a higher heparin immobilizing density yielded a lower plasma protein adsorption and lower platelet adhesion on the surface. They also pointed out, however, that the hydrophilic PAN-C surface caused more protein adsorption and a reduction of platelet adhesion was not observed. They proposed that a hydrophilic surface did not always correspond to protein resistance and suppression of platelet adhesion. In some cases the proteins themselves were immobilized onto the membrane surface to improve the hemocompatibility. As we known, fibrinogen activates and albumin inhibits the adhesion of platelets (Klee and Hocker 2000). Therefore, albumin was often physically adsorbed or covalently immobilized onto the membrane surface. Liu et al. (2005) studied the hemocompatibility of PAN membranes covalently immobilized with either collagen (denoted as PAN-PC), which promoted platelet adhesion, or HSA (denoted as PAN-PA), which inhibited platelet adhesion. It can be seen from Table 2 that both platelet adhesion and HPF adsorption were promoted on PAN-PC membrane even compared with original PAN membrane. Both the platelet adhesion and the HPF adsorption of PAN-PA membranes stood almost at the same level with those of the PAN-H membrane. They found that immobilization with HSA resulted in a rougher surface morphology of the PAN membrane, reduced platelet adhesion, fibrinogen adsorption, prolonged the blood coagulation times, and reduced leukopenia and anaphylatoxin. Collagen-immobilized PAN membranes, however, exhibited the opposite effect when contacting blood, although they induced the least complement activation. They indicated that not all plasma proteins are capable of improving hemocompatibility.

262

Fig. 14. Chemical scheme of the immobilization of heparin or insulin. DMF Dimethylformamide

L.-S. Wan et al.


263

We also immobilized heparin or insulin onto the PANCMA membrane; the procedure is illustrated in Fig. 14. Through the platelet- and macrophage-adhesion experiments, it was found that the immobilization of heparin or insulin improved the hemocompatibility of PANCMA membranes.

10.6 Biomimetic Modification A potential technique for reducing protein adsorption on polymeric membranes is to mimic a biologic surface in nature, commonly named biomimetic modification. For example, the red blood cell plasma membrane, unlike synthetic polymer membranes, naturally resists protein fouling. This property may be attributed to the unique phospholipid bilayer structure of the biomembrane. Therefore, great efforts have been made to immobilize phospholipid molecules onto the polymer surface to build a biocompatible surface. To improve the physical and chemical stability, phospholipid molecules with polymerizable groups were synthesized. Researchers have synthesized many kinds of phospholipid-analogous polymers and developed various methods to immobilize them. We have developed a PAN-based membrane containing phospholipid moieties to improve its protein resistance and hemocompatibility. Two methods were adopted to build the biomimetic surface containing phospholipid moieties. As illustrated in Figs. 15 and 16, one is copolymerizing acrylonitrile with a polymerizable monomer containing phospholipid moieties, the other is to introduce phospholipid moieties onto the acrylonitrile/HEMA copolymer (PANCHEMA) membrane surface by chemical reactions. The resultant membrane containing phospholipid moieties was denoted as a PMANCP membrane. Results from Fourier transfer infrared spectroscopy, x-ray photoelectron spectroscopy, and proton/phosphor nuclear magnetic resonance confirmed the chemical structure of the membrane containing phospholipid moieties. Water contact angle was used to characterize the improvement of hydrophilicity by the introduction of phospholipid moieties. Static contact angles as a function of contact time on the PAN, PANCHEMA, and PMANCP membranes are shown in Fig. 17. It can be seen that the water contact angles on PAN and PANCHEMA membranes decreased slightly with time, whereas those on the PMANCP membrane decreased sharply. Furthermore, the water on the PMANCP membrane surface extended out in a few minutes. This was due to the interaction between the water and the polar group of zwitterion moieties on the membrane surface. Advancing and receding water contact angles are presented in Fig. 18. As reported by Ulbricht et al. (1998), grafting hydrophilic polymers on the PAN-based membrane surface only

264

L.-S. Wan et al.

Fig. 15. Schematic representation of the synthesis of PAN containing phospholipid moieties. THF Tetrahydrofuran

had a slight effect on the contact angle; similar results were obtained from the measurements. However, the hydrophilicity was effectively improved by introducing the phospholipid moieties onto the PAN-based membrane, it can be seen from Fig. 18 that the dynamic contact angles on the phospholipid-moiety-modified membranes were obviously lower than on the PANCHEMA membranes. According to the platelet-adhesion experiments, the hemocompatibility of the PAN membrane was remarkably improved by the introduction of phospholipid moieties. This was also confirmed by the static protein adsorption, which is shown in Fig. 19. It was found that the amount of BSA adsorbed on the PAN and PANCHEMA membranes increased almost linearly with the increase in the concentration of BSA. However, differing


265

Fig. 16. Schematic representation for introducing phospholipid moieties onto the acrylonitrile/2-hydroxyethyl methacrylate (HEMA) copolymer (PANCHEMA) membrane surface

Fig. 17. Water contact angle as a function of contact time on the PAN-based membranes. The HEMA mole fraction in the PANCHEMA membrane is 0% (a), 6.4% (b), 9.3% (c), and 17.8% (d). The mole fraction of phospholipid moiety on the PANCHEMA membrane containing phospholipids (PMANCP) membrane surface is 6.09% (e), 9.19% (f), and 17.1% (g)

266

L.-S. Wan et al.

Fig. 18. Static (SCA, black columns), advancing (ACA, hatched columns), and receding (RCA, white columns) contact angles on the PAN-based membranes. The HEMA mole fraction in the PANCHEMA membrane is 0% (a), 6.4% (b), 9.3% (c), and 17.8% (d). The mole fraction of phospholipid moiety on PMANCP membrane surface is 6.09% (e), 9.19% (f), and 17.1% (g)

from those on the PAN and PANCHEMA, the adsorbance of BSA onto the PMANCP membranes increased slightly and stayed at a certain low level regardless of further increases in BAS concentration. All samples were immersed in the aqueous medium during the BSA adsorption measurement; in general, the relatively high level of free water fraction on the phospholipid-modified membrane surface might effectively suppress protein adsorption and platelet adhesion. In addition to the aforementioned approaches, other methods such as blending can be used to obtain protein-resistant PAN-based membranes. To sum up, the successful development of a protein-resistant microporous membrane is an interesting issue for both water treatment membranes and blood-contacting membranes.

10.7 Conclusion To some extent, protein resistance is very important for both antifouling of membranes during filtration and improving the hemocompatibility of


267

Fig. 19. BSA adsorption on PAN-based membranes. The HEMA mole fraction in the PANCHEMA membrane is 0% (a), 6.4% (b), 9.3% (c), and 17.8% (d), and the mole fraction of phospholipid moiety on PMANCP membrane surface is 6.09% (e), 9.19% (f), and 17.1% (g)

blood-contact membranes. The methods overviewed above, such as copolymerization, PEG tethering, physical adsorption, biomacromolecule immobilization, and biomimetic modification, provide promising opportunities for protein resistance at the surface of PAN-based membranes. Many kinds of modifiers based on different protein–membrane interactions may lead to protein resistance in the same way. Among them, hydrophilic, charged, sterically hindered, or biomimetic agents ought to receive preference. Since there are so many modification methods and each approach possesses innate relative merits, one should choose the most suitable method in terms of the given aim and the varying membrane system. Acknowledgements. Financial support obtained from the National Natural Science Foundation of China (Grant no. 50273032) and the National Basic Research Program of China (Grant no. 2003CB15705) are gratefully acknowledged. The authors thank Dr. Fu-Qiang Nie and Dr. Rui-Qiang Kou very much for their contribution.

268

L.-S. Wan et al.

References Dee KC, Puleo DA, Bizios R (2002) An Introduction to Tissue–Biomaterial Interactions. John Wiley & Sons, New York Etheve J, Dejardin P, Boissiere M (2003) Adsorption of lysozyme on a hemodialysis sulfonated polyacrylonitrile membrane, with and without preadsorbed poly(ethyleneimine) on the external faces. Colloids Surf B 28:285–293 Godjevargova T, Konsulov V, Dimov A, Vasileva N (2000) Behavior of glucose oxidase immobilized on ultrafiltration membranes obtained by copolymerizing acrylonitrile and N-vinylimidazol. J Membr Sci 172:279–285 Groth T, Seifert B, Malsch G, Albrecht W, Paul D, Kostadinova A, Krasteva N, Altankov G (2002) Interaction of human skin fibroblasts with moderate wettable polyacrylonitrilecopolymer membranes. J Biomed Mater Res 61:290–300 Hayama M, Yamamoto K, Kohori F, Sakai K (2004) How polysulfone dialysis membranes containing polyvinylpyrrolidone achieve excellent biocompatibility? J Membr Sci 234:41–49 Huang XJ, Xu ZK, Wan LS, Wang ZG, Wang JL (2005) Novel acrylonitrile-based copolymers containing phospholipid moieties: synthesis and characterization. Macromol Biosci 5:322–330 Huang XJ, Xu ZK, Wan LS, Wang ZG, Wang JL (2005) Surface modification of polyacrylonitrile-based membranes by chemical reactions to generate phospholipid moieties. Langmuir 21:2941–2947 Klee D, Hocker H (2000) Polymers for biomedical applications: improvement of the interface compatibility. Adv Polym Sci 149:1–57 Krasteva N, Harms U, Albrecht W, Seifert B, Hopp M, Altankov G, Groth T (2002) Membranes for biohybrid liver support systems – investigations on hepatocyte attachment, morphology and growth. Biomaterials 23:2467–2478 Lin WC, Liu TY, Yang MC (2004) Hemocompatibility of polyacrylonitrile dialysis membrane immobilized with chitosan and heparin conjugate. Biomaterials 25:1947–1957 Liu TY, Lin WC, Huang LY, Chen SY, Yang MC (2005) Hemocompatibility and anaphylatoxin formation of protein-immobilizing polyacrylonitrile hemodialysis membrane. Biomaterials 26:1437–1444 Musale DA, Kulkarni SS (1996) Fouling reduction in poly(acrylonitrile-co-acrylamide) ultrafiltration membranes. J Membr Sci 111:49–56 Musale DA, Kulkarni SS (1997) Relative rates of protein transmission through poly(acrylonitrile) based ultrafiltration membranes. J Membr Sci 136:13–23 Nie FQ, Xu ZK, Huang XJ, Ye P, Wu J (2003) Acrylonitrile-based copolymer membranes containing reactive groups: Surface modification by the immobilization of poly(ethylene glycol) for improving antifouling property and biocompatibility. Langmuir 19:9889– 9895 Nie FQ, Xu ZK, Ming YQ, Kou RQ, Liu ZM, Wang SY (2004) Preparation and characterization of polyacrylonitrile-based membranes: effects of internal coagulant on poly(acrylonitrile-co-maleic acid) ultrafiltration hollow fiber membranes. Desalination 160:43–50 Nie FQ, Xu ZK, Qian Y, Jian W, Wan LS (2004) Surface modification of poly(acrylonitrileco-maleic acid) membranes by the immobilization of poly(ethylene glycol). J Membr Sci 235:147–155 Nie FQ, Xu ZK, Wan LS, Ye P, Wu H (2004) Acrylonitrile-based copolymers containing reactive groups: synthesis and preparation of ultrafiltration membranes. J Membr Sci 230:1–11 Nie FQ, Xu ZK, Ye P, Wu J, Seta P (2004) Acrylonitrile-based copolymer membranes containing reactive groups: effects of surface-immobilized poly(ethylene glycol)s on antifouling properties and blood compatibility. Polymer 45:399–407


269

Sun SD, Yue YL, Huang XH, Meng DY (2003) Protein adsorption on blood-contact membranes. J Membr Sci 222:3–18 Thomas M, Vakette P, Mausset A-L, Déjardin P (2000) High molecular weight kininogen adsorption on hemodialysis membranes: influence of pH and relationship with contact phase activation of blood plasma. Influence of pre–treatment with poly(ethyleneimine). Inter J Artif Organs 23:20–28 Ulbricht M, Belfort G (1996) Surface modification of ultrafiltration membranes by low temperature plasma II. Graft polymerization onto polyacrylonitrile and polysulfone. J Membr Sci 111:193–215 Ulbricht M, Richau K, Kamusewitz H (1998) Photomodification of ultrafiltration membranes – Part 11 – Chemically and morphologically defined ultrafiltration membrane surfaces prepared by heterogeneous photo-initiated graft polymerization. Colloids Surf A 138:353–366 Valette P, Thomas M, Dejardin P (1999) Adsorption of low molecular weight proteins to hemodialysis membranes: experimental results and simulations. Biomaterials 20:1621– 1634 Wan LS, Xu ZK, Huang XJ, Wang ZG, Wang JL (2005) Copolymerization of acrylonitrile with N-vinyl-2-pyrrolidone to improve the hemocompatibility of polyacrylonitrile. Polymer 46:7715–7723 Xu ZK, Kou RQ, Liu ZM, Nie FQ, Xu YY (2003) Incorporating alpha-allyl glucoside into polyacrylonitrile by water-phase precipitation copolymerization to reduce protein adsorption and cell adhesion. Macromolecules 36:2441–2447 Xu ZK, Nie FQ, Qu C, Wan LS, Wu J, Yao K (2005) Tethering poly(ethylene glycol)s to improve the surface biocompatibility of poly(acrylonitrile-co-maleic acid) asymmetric membranes. Biomaterials 26:589–598 Xu ZK, Qian Y, Kou RQ, Jian W, Wang JQ (2004) First results of hemocompatible membranes fabricated from acrylonitrile copolymers containing sugar moieties. J Membr Sci 243:195–202 Yan F, Dejardin P, Mulvihill JN, Cazenave J-P, Crost T, Thomas M, Pusineri C (1992) Influence of a preadsorbed terpolymer on human platelet accumulation, fibrinogen adsorption, and ex vivo blood activation in hemodialysis hollow fibers. J Biomater Sci Polym Ed 3:389–402

11

Modulation of the Adsorption and Activity of Protein/Enzyme on the Polypropylene Microporous Membrane Surface by Surface Modification Qian Yang, Zhi-Kang Xu, Zheng-Wei Dai

Abstract. As an excellent membrane material, polypropylene microporous membrane (PPMM), has received much consideration in recent years. PPMM has also been used widely in many fields such as water treatment, enzyme immobilization, and blood oxygenation. However, poor surface properties cumber the further applications of this material. Many efforts have been made to solve this problem, and surface modification seems to be the most efficient way. In this article we introduce some work in which surface modification was conducted to reduce nonspecific protein adsorption and to modulate the activity of immobilized enzymes on the PPMM surface. We focus on the studies of our group, although other researches are also discussed.

11.1 Surface Modifications for Reducing Nonspecific Protein Adsorption There have been many interests in membrane processes for last decade because of they are extremely efficient, have a low energy consumption, and are easy to carry out. Nowadays, membranes are used widely, especially in biomedical applications such as dialysis, plasmapheresis, and oxygenation of blood during cardiac surgery. However, it is well known that the major obstacle to the extensive use of membrane processes in therapeutic treatment is protein fouling of polymeric membrane materials. Protein deposition on the membrane surface can cause unstable transport characteristics, and cellular interactions with artificial surface are also assumed to be mediated through adsorbed proteins (Deppisch et al. 1998). Designing a polymer surface that rejects proteins (i. e., a nonfouling surface) has been a central issue in the field of biomedical materials research (Ikada 1994; Klee and Hocker 1999; Ratner et al. 1979). The adsorption of proteins is highly complex. We do understand that these interactions are determined by the Qian Yang, Zhi-Kang Xu, Zheng-Wei Dai: Institute of Polymer Science, Zhejiang University, Hangzhou 310027, People’s Republic of China, E-mail: [email protected] Principles and Practice Proteins at Solid–Liquid Interfaces Philippe Déjardin (Ed.) © Springer-Verlag Berlin Heidelberg 2006

272

Q. Yang et al.

hydrophobic/hydrophilic, charged/uncharged, and polar/nonpolar parts of the proteins, and by the nature of the membrane surface. The adsorbed protein film shows time-dependent conformational changes, which may cause desorption or protein exchange. Adsorption processes are described by the typical Langmuir isotherms. After a long contact time, a stationary state is reached, which corresponds to an irreversible protein adsorption (Lundstrom and Elwing 1990; Sonderquist and Walton 1980). However, in spite of extensive investigations (Guell et al. 1999; Kuberkar and Davis 2000; Mueller and Davis 1996; Tie et al. 2003), the mechanism of protein adsorption onto the membrane surface remains unsolved. Nevertheless, The primary method of reducing protein adsorption onto polymeric materials is surface modification. There are many methods of surface modification, as outlined in Table 1 (Ratner 1995). Among them, the covalent method offers a more stable modification surface as compared to other methods (Gupta and Anjum 2003). Graft polymerization is one such method in which polymer chains are tethered to the material surface. Grafting has several advantages over other methods, including easy and controllable introduction of graft chains with a high density and exact Table 1. Physical and chemical surface modification methods. UV Ultraviolet, RF radio frequency Covalently attached coatings Radiation grafting (electron accelerator and gamma) Photografting (UV and visible sources) Plasma (gas discharge; RF, microwave, acoustic) Gas phase deposition Ion beam sputtering Chemical vapor deposition Chemical grafting (e. g., ozone treatment + grafting) Silanization Biological modification (biomolecule immobilization) Modification of the original surface Ion beam etching (e. g., argon, xenon) Ion beam implantation (e. g., nitrogen) Plasma etching (e. g., nitrogen, argon, oxygen, water vapor) Corona discharge (in air) Electron beam treatment Ion exchange UV irradiation Chemical reaction Nonspecific oxidation (e. g., ozone) Functional group modifications (oxidation, reduction) Addition reactions (e. g., acetylation, chlorination) Conversion coatings (phosphating, anodization)

11 Modulation of Adsorption and Activity of Protein/Enzyme on PPMM Surface 273

location of graft chains to the surface, with the bulk properties remaining unchanged. This method is also applied in membrane surface modification to reduce protein adsorption and there are many research groups making efforts to develop membrane surface biocompatibility using this method. Hydrophilic polymers are often used for this purpose because of their wettability and biocompatibility. However, derivatives of native substances existing in the biological systems (such as phospholipid, carbohydrate, and polypeptide), which have excellent biocompatibility, are thought to be efficient for the reduction of protein deposition. Herein, the discussion will focus on these biomimetic polymer modifiers used by our group, although other polymers will also be mentioned.

11.1.1 Plasma treatment Plasma modification of polymeric materials is an extremely useful way of tailing a polymer into a desired material by utilizing the selective chemistry and molecular structure on the surface (Kiaei et al. 1995; Lee et al. 1991; Oehr et al. 1999). It is an efficient way to produce functional groups such as hydroxyl groups, amino groups, and carboxylic groups. Plasma can also initiate polymerization reactions of monomers on a surface. Graft polymerization can be carried out by plasma treatment either directly or indirectly. For the former, monomers are exposed to the plasma environment on the objective surface. At the same time, indirect polymerization can also be initiated by functional groups produced by the plasma. Kou et al. (2003) grafted a glycopolymer to polypropylene microporous membrane (PPMM) surface by direct plasma polymerization (Fig. 1a). A monomer containing sugar moieties, α-allyl glucoside (AG), was synthesized and grafted to PPMMs (Fig. 1b). In that case the PPMM was immersed in an AG solution of N,N-dimethylformamide for a predetermined time, and the solvent was evaporated in a vacuum oven. The coated AG monomer was then grafted chemically onto the membrane surface by N2 plasma. Bovine serum albumin (BSA) was used as a model protein to evaluate the anti-protein-fouling characteristic of this modified PPMM. As shown in Fig. 2, increases in the AG grafting degree leads to a reduction in the amount of protein deposited to the membrane surface. This is due to the hydrophilicity and biocompatibility of the grafted AG. Liu et al. (2003) tethered a polypeptide, poly(γ -stearyl-L-glutamate) (PSLG), to the PPMM surface by plasma treatment (Fig. 3). First, the PPMM was treated by plasma under an atmosphere of ammonia to introduce amino groups to the surface. Then, a monomer, N-carboxyanhydride (NCA), was subjected to ring-opening polymerization initiated by the amino groups on

274

Q. Yang et al.

Fig. 1. Molecular structure of α-allyl glucose (AG; a) and poly (α-allyl glucoside) (PAG)modified polypropylene microporous membrane (PPMM; b)

Fig. 2. a–d Effect of AG grafting degree on bovine serum albumin (BSA) adsorption (a–d: 0, 0.82, 1.86, 3.46 wt.%, respectively)

the membrane surface, and PSLG was grafted onto the membrane surface. There are two mechanisms of ring-opening polymerization of NCA (Fig. 3b) and the active monomer mechanism, which is initiated by hydroxyl groups, must be depressed to improve the grafting degree. Thus γ -(aminopropyl) triethanoxysilane (γ -APS) was used to eliminate the effect of hydroxyl groups by reacting with them. The results of BSA adsorption experiments are shown in Fig. 4.


Fig. 3. Poly(γ -stearyl-L-glutamate) (PSLG)-modified PPMM (a) and mechanisms for the synthesis of polypeptide from N-carboxyanhydride (NCA) monomers (b)

276

Q. Yang et al.

Fig. 4. BSA adsorption onto different PPMMs. original membrane; NH3 -plasma-treated membrane; PSLG-grafted membrane without γ -(aminopropyl) triethanoxysilane (γ -APS) treatment; PSLG-grafted membrane with γ -APS treatment

For the ammonia-plasma-treated membranes, BSA adsorption was reduced slowly as a result of the hydrophilic groups generated on the membrane surface. However, what surprised us was that the PSLG-γ -APS-PPMM exhibited greatly increased protein adsorption, even larger than the nascent PPMM. This could be interpreted by the conformation of PSLG on the surface. It is known that polypeptide exhibits α-helix and coil conformations under different conditions. For the poly(γ -stearyl-L-glutamic acid) with long stearyl groups, Poche et al. (1995) speculated that the grafted PSLG in solutions formed an α-helix conformation that was supported by the intramolecular hydrogen bonds, with the peptide main chains in the cores and γ -stearyl long side chains stretched outside. This molecular model of PSLG could be used to explain the BSA adsorption results. For the PSLG-PPMM, the amount of adsorbed BSA increased a little; this could be ascribed to the existence of -OH groups, which led to low polymerization degree. For the PSLG-γ -APS-PPMM, the polymerization degree of grafted chains increased, the stearyl long side chains stretched outside, thus greatly increasing the surface hydrophobicity, resulting in an increase in BSA adsorption. This characteristic was used to immobilize lipase in our further work.

11.1.2 Ultraviolet (UV) modification UV irradiation is an effective technique for membrane surface modification. It may be the most appropriate of all surface modification techniques


because of certain features: the low cost of operation, mild reaction conditions, it is highly surface-selective, and because it alters the material surface with facile control of the chemistry (Ma et al. 2000a). Several studies have been carried out using photoinitiators (Kita et al. 1994; Richey et al. 2000; Ulbricht et al. 1998) or specially synthesized molecules with UV-sensitive end groups (Taniguchi et al. 2003; Thom et al. 2000). Figure 5 shows a typical two-step process for UV-induced polymerization (Ma et al. 2000b). Xu and coworkers (2004) modified PPMMs using phospholipid-analogous polymers (PAPs) (Fig. 6). N,N-dimethylaminorthyl methacrylate (DMAEMA) was grafted to the membrane surface by photoinduced graft polymerization and then reacted with 2-alkyloxy-2-oxo-1,3,2-dioxaphospholanes. Five 2-alkyloxy-2-oxo-1,3,2-dioxaphospholanes, containing octyloxy, dodecyloxy, tetradecyloxy, hexadecyloxy, and octadecyloxy groups in their molecular structure, were used to fabricate the PAP-modified PPMMs. The influence of grafting degree on BSA adsorption is shown in Fig. 7. It can be seen that the adsorbed BSA on the membrane decreased with increasing grafting degree for both poly(DMAEMA)-grafted and PAPmodified membranes at the studied range. The five PAP-modified membranes had better protein resistance than the unmodified polypropylene membrane and poly(DMAEMA)-grafted membranes. Nevertheless, as the number of carbon atoms was increased from 14 to 16 and 18, BSA adsorption on the membrane surface was clearly suppressed. The reason for this was probably the mimetic characteristics of the PAP-modified membranes; the principal components of a biomembrane are lipids, proteins, and carbo-

Fig. 5. Schematic diagram of a novel two-step process for photoinduced graft polymerization

278

Q. Yang et al.

Fig. 6. Schematic representation for the fabrication of phospholipid-analogous polymer (PAP)-modified PPMMs. DMAEMA N,N-dimethylaminorthyl methacrylate

Fig. 7. BSA adsorption on poly(DMAEMA)-grafted and PAP-modified membranes


hydrates. However, the amount of proteins and carbohydrates is relatively small. Most of the lipids are phospholipids such as phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, phosphatidic acid, phosphatidylinositol, and phosphatidylglycerol. As part of their structure, these phospholipids all have two long alkyloxy groups. Two layers of phospholipid molecules are facing each other, burying the hydrophobic moieties inside the membrane. The zwitterionic species cover the membrane surface and there is a reorientation of the PAP to put the zwitterions at the surface of PAP-modified PPMMs. During the protein adsorption measurements, the PAP-modified PPMMs were immersed in BSA solution for 24 h. In that case, the membranes were in an aqueous environment and should present the zwitterions at the surface. Thus, in these conditions, the membrane surfaces were hydrophilic and were relatively similar to those of the biomembranes. The hydrophilic surfaces normally facilitate the reduction of protein adsorption on the membrane (Ma et al. 2000a; Steen et al. 2002; Wavhal and Fisher 2003; Xu et al. 2003). Furthermore, with an increase of the number of carbon atoms of the long alkyloxy groups, the impact of the original membrane surface on protein adsorption could be reduced, which leads to a decrease in protein adsorption on the PAP-modified membranes. Therefore the membranes exhibit excellent protein resistance. N-Vinyl-2-pyrrolidone is a hydrophilic and nonionic monomer that can be easily initiated through radical, thermal, or photo irradiation (Senogles and Thomas 1975). Poly(N-vinyl-2-pyrrolidone) (PVP) is a polymer with great potential applications in different areas of biomedicine due to its ex-

Fig. 8. Molecular structure of N-vinyl-2-pyrrolidone (NVP) and ultraviolet (UV)-induced grafting polymerization

280

Q. Yang et al.

Fig. 9. a Effect of grafting degree (GD) on BSA adsorption in PVP-modified PPMMs. • nascent PPMM; PVP-modified PPMM GD = 1.49 wt. % and PVP-modified PPMM GD = 4.08 wt. %. b Static contact angle of PVP-modified PPMMs with different grafting degrees


cellent biocompatibility with living tissues and extremely low cytotoxicity (Wetzels and Koole 1999). Due to its outstanding performance, PVP has also been used widely as a surface-modification agent. Liu et al. (2004) used PVP as modifier to improve the surface properties of PPMMs by UVinduced graft polymerization of N-vinyl-2-pyrrolidone, as shown in Fig. 8. After grafting PVP to the membrane surface, as been expected, protein adsorption was markedly reduced with increasing grafting degree (Fig. 9). Usually, materials processing hydrophilic surface show relatively low nonspecific adsorption for proteins or cells. Therefore, the reduction in BSA adsorption could be ascribed mainly to the improvement in hydrophilicity effected by the introduction of PVP chains on the PPMM surface. Yang et al. (2005) grafted a novel glycopolymer containing linear sugar moieties, d-gluconamidoethyl methacrylate (GAMA) (Fig. 10), to the PPMMs by UV-induced graft polymerization and obtained a hydrophilic surface with low protein fouling and relatively high flux recovery (Table 2). After surface modification, the antifouling property was evaluated by filtrating a 1 g/L BSA solution through the membranes. Pure water flux after cleanout was also examined to confirm flux recovery. The ethanol-wetted nascent membrane showed the largest loss of flux within the measurement time, which suggests that a large amount of BSA protein had been deposited on the surface. However, the reduction in flux was suppressed by grafting of GAMA, indicating that the GAMA polymer layer effectively prevented the

Fig. 10. Molecular structure of GAMA Table 2. Permeation and antifouling properties of D-gluconamidoethyl methacrylate (GAMA) grafted polypropylene microporous membrane (PPMMs). JW Flux of pure water, JP flux of 1 g/l bovine serum albumin solution, JR flux after cleanout, RFR relative flux reduction as (1 − JP / JW ) × 100%, FRR flux recovery ratio as JR / JW × 100% Membrane

JW (kg/m2 h) JP (kg/m2 h)

JR (kg/m2 h)

RFR

FRR

Ethanol-wetted 2.23 wt% GAMA grafted 3.5 wt% GAMA grafted 4.58 wt% GAMA grafted 5.47 wt% GAMA grafted 6.03 wt% GAMA grafted

340 ± 7 466±26 608±21 729±14 762±22 764±28

219 ± 22 390±26 488±38 646±30 658±33 649±35

73 60 57 38 57 59

64 84 80 89 86 85

93 ± 22 187±24 348±37 452±38 434±32 452±25

282

Q. Yang et al.

adsorption of BSA. Moreover, the recovery flux increased significantly with increasing grafting degree, and a relatively high flux recovery ratio (> 80%) was achieved even at lower grafting degree. All of these improvements in surface properties can be ascribed to the highly hydrophilic nature of the linear sugar moieties in the grafted chains.

11.1.3 γ -Ray-induced modification

γ -Ray-induced graft polymerization has been used extensively because it

is the most versatile and promising one due to its rapid formation of active sites on the substrate surface and in the material matrix. Almost all of the polymeric materials can be modified by γ -ray with different monomers and the resulting material is highly pure as no initiator or related impurities remain in the matrix. However, as a very high energy radiation, γ -rays have a tremendous penetrability such that active sites are also generated in the material bulk, which often affects the inherent bulk properties. Table 3. Characteristics and bovine serum albumin (BSA) permeation properties of 2-hydroxyethyl methacrylate (HEMA)-grafted PPMMs with different grafting degrees (PHn) Sample code PPa PH1 PH2 PH3 PH4 PH5 PH6 a b c d e f g h I j

Flux Resistance Degree (1014 m−1 ) of graftingb (l/m2 h) (wt %) Rc g Jw c Js d Rm e Rt f 2.9 7.2 12.5 24.0 32.2 40.6

77.1 75.2 69.9 68.6 68.1 66.3 60.9

24.2 26.9 27.8 28.7 37.6 43.6 44.8

4.67 4.78 5.15 5.28 5.43 5.64 5.91

Surface Adsorbed Wettabilityh potential BSAj (cos θ) ζi (mV) (µg/cm2 )

14.8 10.1 −0.32 13.4 8.60 −0.15 12.9 7.75 0.03 12.5 7.22 0.27 9.57 4.14 0.31 8.25 2.61 0.44 8.03 2.12 0.56

−36.47 −31.33 −29.10 −26.72 −11.57 −7.75 −5.22

104.5 83.6 69.4 38.1 20.9 26.9 9.0

PPMM with exposure of γ -rays Degree of grafting (wt %) = (wg − w0 )/ w0 × 100, where w0 and wg are the weights of the membrane before and after the grafting reaction, respectively Deionized water flux Flux of 1 g/l BSA in 10 mM PBS of pH 7.4 at 23 ± 2 ◦ C The membrane resistance (Rm ) was calculated by the measured deionized water flux The total resistance (Rt ) during the filtration of protein solution was calculated by the flux (Js ) of BSA protein solution Rc caused by BSA protein in solution was calculated by subtracting Rm from Rt The average estimated error for cosθ was ±0.04 The average estimated error was ±9% Irreversible adsorptive fouling of the membrane was defined as the amount of BSA absorbed onto the surface of the membrane after chemical cleaning


Kang et al. (2001) used Co60 as a γ -ray source for surface modification on the PPMMs by graft polymerization of 2-hydroxyethyl methacrylate (HEMA). Through adjusting the radiation dose, modified PPMMs with different grafting degrees were prepared. Protein adsorption measurements were carried out on these membranes by filtration of BSA solution through the membranes. The results are shown in Table 3 and Fig. 11. The flux of deionized water decreased with increases in the grafting degree of polyHEMA (PHEMA). On the other hand, the flux of BSA buffer solution increased with increasing PHEMA grafting degree. Clearly, the more hydrophilic the surface (increasing cosθ value), the lower the flux decline with

Fig. 11. Buffer flux as a function of the 2-hydroxyethyl methacrylate (HEMA) grafting degree

Fig. 12. Schematic illustration of ozone-induced graft polymerization

284

Q. Yang et al.

BSA buffer solution. The decrease of flux of the deionized water could be explained by narrowed or plugged pores as a result of the swelling of grafted PHEMA in a buffer solution. In the case of the permeation of BSA buffer solution, although the pores narrowed with increasing PHEMA grafting degree, the increased flux of the PHEMA-grafted membranes was due to the weakening of hydrophobic interaction between the BSA molecules and the hydrophilic membrane surface. Therefore, the flux of BSA buffer solution increased because the fouling caused by the adsorption of BSA was reduced with increasing PHEMA grafting degree.

Fig. 13. a Amount of protein adsorbed per unit area of the virgin PPMM, grafted (ozone treated for 1, 3, and 5 min) and the commercial modified GVHP (hydrophobic) and GVWP (hydrophilic) membranes. b Fluxes for virgin and modified membranes during BSA solution permeation


11.1.4 Ozone Method Wang et al. (2000) performed surface modification of PPMMs by using ozone to introduce peroxide onto the membrane surface. After that, graft polymerization of HEMA was carried out. The polymerization was initiated at a mild temperature by redox decomposition of the peroxide (Fig. 12). Similar to the results of Kang et al. (see previous section), these HEMAgrafted PPMMs also showed great protein resistance (see Figs. 13a,b and 14). The trend of flux decline for the modified (grafted) membrane was nearly the same as that of the unmodified (virgin) membrane. However, the flux recovery efficiency (Fig. 14), which is closely associated with the reversibility of membrane fouling layers, increased markedly with increasing ozone-treatment time by up to 5 min. It appears that membrane fouling might be the result of accumulation of protein aggregates at the membrane surface regardless of the membranes used (grafted or not), but could be reversed more easily for the grafted membranes with an appropriate time of ozonation. This can be attributed to the more hydrophilic nature of the modified membrane surfaces. Also, the steric hindrance resulting from the grafted PHEMA chains may play an important role in preventing the direct interaction between the membrane surfaces and proteins.

Fig. 14. Flux recovery efficiencies with water flushing at the end of operation

286

Q. Yang et al.

11.2 Surface-Modified PPMMs for Enzyme Immobilization Catalysis is frequently a prerequisite for efficiency in organic reactions (Woltinger et al. 2001). As biocatalysts, enzymes exhibit several features that make their use advantageous compared to conventional chemical catalysts. The specific features presented by enzymes are the high level of catalytic efficiency and high degree of specificity including substrate specificity, region specificity and stereospecificity. However, conventional catalytic processes are often carried out in a homogeneous way, which frequently renders the separation of the catalyst from the reaction medium a cumbersome nuisance (Annis and Jacobsen 1999; Garber et al. 2000). Moreover, the loss of the precious enzymes in these catalysis processes becomes a great obstacle for commercial application. Therefore, efforts have been made to bind enzymes to an insoluble support/carrier. In recent years, artificial membranes have been applied in biotechnology because of their interesting properties of high specific surface area and the possibility of combining separation with the chemical reaction (Gekas 1986). Among these membranes, the polypropylene membrane is particularly interesting due to its well-controlled porosity, chemical and thermal inertness, and high potential for comprehensive applications. However, PPMMs are pressed for polar functional groups, which are necessary for the covalent immobilization of proteins. Moreover, the poor biocompatibility of this membrane may cause nonbiospecific interactions, protein denaturation, and loss of enzyme activity (Kasemo 2002). Thus, one can envisage that it is possible to introduce a reactive and biofriendly interface on the PPMM surface for enzyme immobilization through surface modification technologies, which may reduce some of the nonspecific enzyme–support interactions, create a specific microenvironment for the enzyme, and benefit the enzyme activity (Deng 2004b). The introduction of polymer chains carrying polar and reactive groups to the PPMM surface may solve this problem easily. Tethering of polymers to the membrane surface can offer relatively high functional group density and depress the hindrance effect, which may cumber the immobilization of the enzyme. In general, the polymer carboxylic, amino, or thio groups, grafted to the membrane surface, are used for the immobilization of enzymes (Sano et al. 1993). Hydrophilic monomers, such as acrylic acid, HEMA, glycidyl methacrylate (GMA), are often used to modify the membrane surface, enhancing the hydrophilicity and biocompatibility of the membranes. However, for some enzymes, hydrophilic polymers are not always a suitable choice. In the case of lipase, the enzyme is activated in the presence of aqueous–hydrophobic interfaces. In addition, polymers that exhibit semiconductivity are preferred for the immobilization of enzymes that catalyze reactions via electron trans-


fer. Anyway, the purpose of modification is to provide the enzyme with conditions similar to those that it requires in nature. There are three methods for immobilizing enzymes: physical adsorption/entrapment, direct grafting, and site-specific immobilization. In general, physical adsorption/entrapment offers better enzyme activity but relatively low stability. On the contrary, covalent immobilization is much more stable but often causes the enzyme to denature.

11.2.1 Physical Adsorption/Entrapment Physical adsorption is the most convenient way to achieve enzyme immobilization, avoiding the tedious and time-consuming multistep procedure of covalent binding (Aleixo et al. 1985; Cleveland et al. 1981). Entrapment immobilizes enzymes by trapping them into porous materials (membranes). Because there is no covalent attachment, the activity of the enzymes immobilized by physical adsorption, and entrapment is easily maintained. However, loss of enzyme in applications is inevitable for the physical adsorption method, impeding the further utility of this method. At the same time, how substrates diffuse into the porous materials and how products diffuse out are the critical problems associated with the entrapment strategy. Deng et al. (2004a,b,c) immobilized Candida rugosa lipase on a series of surface-modified PPMMs by adsorption and compared their effects on the enzyme activity. Three different kinds of modifiers, poly(α-allyl glucoside) (PAG), two polypeptides with short and long hydrophobic side chains, poly(γ -ethyl-L-glutamate) and poly(γ -stearyl-L-glutamate), and phospholipid-analogous polymers (PAP) containing hydrophobic octyloxy, dodecyloxy, and octadecyloxy groups (8-PAP, 12-PAP, and 18-PAP respectively), were tethered onto PPMMs (see Figs. 1, 3, and 6). Then lipases from Candida rugosa were immobilized on these membranes by adsorption and their activity was examined. The specific activity and the activity retention of the lipase immobilized onto the hydrophobic polypeptide- and PAP-modified membranes were both higher than those of nascent and hydrophilic PAG-modified membranes (Table 4). This can be ascribed to the large hydrophobic surface that surrounds the catalytic site of the lipase; lipase is thought to be activated in the presence of a hydrophobic interface, which can be helpful for the rearrangement of the protein conformation to yielding an “open state” of the lipase active site. There is a strong hydrophobic interaction between the long alkyl chain of PAP and polypeptides and the hydrophobic domain around the active site of lipase, which stabilizes the “open state” conformation of lipases and favors the accessibility of the

288

Q. Yang et al.

Table 4. Comparing of lipase activity as function of different modifier. PAG Poly (α-allyl glucoside), PAP phospholipids analogous polymers, PELG poly(γ -ethyl-L-glutamate), PSLG poly(γ -stearyl-L-glutamate)

Membrane/Modifier Nascent PPMM PAG

PAP Polypeptide

8-PAP 12-PAP 18-PAP PELG PSLG

Grafting degree (wt.%) – 2.1 2.8 3.7 10.6 10.9 10.8 3.5 3.6

Water contact Specific activity (U/mg protein) angle (◦ C) 118 69.9 60 61.3 48 62.3 36 60.7 53 90.1 92 94.3 96 101.2 113 76.4 122 88.1

Activity retention (%) 57.5 ± 2.8 50.4 ± 2.7 51.2 ± 2.8 49.9 ± 2.9 74.1 ± 3.2 77.5 ± 3.7 83.2 ± 3.3 62.8 ± 3.3 72.4 ± 3.9

active site to substrates. On the other hand, after modification with PAG, the sharp increase in hydrophilicity induces the conformational equilibrium of lipase, to some extent shifting it toward the unfavorable “closed state”. Tanioka et al. (1998) entrapped invertase in the pores of PPMMs, and these membranes were used to hydrolyze sucrose. As show in Fig. 15, they embedded invertase into the pores of PPMMs and then grafted poly(acrylic acid) (PAA) onto the membrane surface to entrap the enzyme. The invertase-entrapped PPMM was immersed in sucrose aqueous solution to hydrolyze the sucrose. The sucrose that permeated through the PAA layer on the membrane surface was hydrolyzed into fructose and glucose by invertase, and the decomposed products exited to the external solution. The hydrolyzation process of the sucrose caused a change in optical rotation, and inspection could be made using a polarimeter. Figure 16a,b shows the hydrolysis process of the invertase-immobilized PPMMs with a high (Fig. 16a) and low (Fig. 16b) PAA grafting degree, respectively, as a function of time. Arrows indicate that the membranes were removed from the sucrose solution. After this removal, hydrolysis cannot be observed at all in the case with the high grafting degree (Fig. 16a), but can still be observed in that of the low grafting degree (Fig. 16b). This indicates that invertase leakage occurred in the case of a low grafting degree, in which the PAA layer was not thick enough to prevent the invertase from spilling. However, with higher PAA grafting degree invertase could be entrapped stably in the PPMM pores and the hydrolysis processed very well.

11 Modulation of Adsorption and Activity of Protein/Enzyme on PPMM Surface 289 Fig. 15. Schematic representation of enzyme entrapment

11.2.2 Covalent Binding Covalent binding means that the enzyme is immobilized by a chemical reaction between the enzyme and the support material to form covalent bonds. Although this immobilization generally enhances enzyme stability, one major disadvantage of this method is that the activity of the immobilized enzyme is often significantly decreased because the active site may be blocked, multiple point-binding may occur, or the enzyme may be denatured (Butterfield et al. 1994; Ganapathi et al. 1998; Zhuang and Butterfield 1992, 1993). As shown in Fig. 17a, the enzyme is directly immobilized, often through the ε-amino functionality of lysine residues on the protein. Because the protein often contains multiple lysine residues, which spread over the enzyme surface, different orientations of the enzyme occur, and it may thus be difficult for the active site of the enzyme to interact with the substrate.

290

Q. Yang et al.

Fig. 16. Hydrolysis process of high (a) and low (b) poly(acrylic acid) grafting degree


Fig. 17. Schematic representation of direct graft (a) and site-specific (b) immobilization of enzymes

Piletsky et al. (2003) functionalized PPMMs with polyaniline (PANI) and immobilized horseradish peroxidase (HRP) onto the membrane surface by both physical adsorption and covalent attachment. PANI is a semiconductive material that is stable in an oxygen atmosphere, contains amino groups, and is suitable for the covalent immobilization of biomolecules through carbodiimide or glutaric dialdehyde chemistry. Electronactive properties of PANI polymer could be envisaged as an additional advantage over conventional polymer matrices for enzyme immobilization, due to the fact that the electron-donating ability of PANI may play an important role in the enzyme catalysis of redox reactions (Malinauskas 1999). The PANI-coated PPMMs showed a high affinity for this enzyme as a result of the combination of electronstatic and hydrophobic interactions between the PANI and proteins (Bossi et al. 2002; Chen et al. 2000a, b). PANI has three oxidation states, leukoemeraldine (reduced), emeraldine, and pernigraniline (oxidized; see Fig. 18), and these different forms possess different adsorption properties to proteins. The highest and lowest bindings were achieved by reduced and oxidized PANI, respectively. HRP immobilized on the PANI-coated PPMM was shown to retain 70% of its activity after 3 months of storage at +5 ◦ C, suggesting that this material can be used for practical application, such as in bioreactors as enzyme membranes. Becker and coworkers (2002) presented a process involving continuous synthesis and simultaneous product release of amylase by an enzymeimmobilized membrane (see Fig. 19). PPMMs were functionalized by photoinitiated graft polymerization of GMA and the enzyme, amylosucrase, was immobilized by the reaction with epoxy groups on the membrane surface. Amylosucrases are found in many Neisseria species and catalyze the reaction sucrose + {1, 4-α-D-glucosyl}n → D-fructose + {1, 4-α-D-glucosyl}n+1 The enzyme transfers the glucose moiety of sucrose to the nonreducing end of an α-1,4-glucan chain, extending it linearly by one glucose unit and thereby releasing a fructose molecule. A sucrose solution and maltooligosaccharide (MOS) mixture containing from 3 to 6 glucose units with

292

Q. Yang et al.

Fig. 18. Oxidation states of polyaniline

Fig. 19. Scheme of enzymatic poly- or oligoglucan synthesis within the membrane pore


α-1,4-linkage was used as a primer to filtrate through the membrane at

37 ◦ C. The formed products (i. e., glucose coupled to the primer molecule), was characterized by high-performance liquid chromatography analysis of the filtrates at different flow rate (Fig. 20). The initial MOS mixture contained oligosaccharides with a degree of polymerization (DP, i. e., number

Fig. 20. Chromatograms of the filtrates with flow rate of 0.2 ml/min. (a) and 0.1 ml/min (b). DP Degree of polymerization

294

Q. Yang et al.

of glucose residues) of 3, 4, 5, and 6, with the DP3 oligosaccharide being most abundant followed by DP4 and DP5 oligosaccharides. At the flow rate of 0.2 ml/min the maltooligosaccharides were enlarged by one glucose unit per membrane passage and a definite shift can be seen in the oligosaccharide distribution: the DP3 peak remains almost constant, the DP4 peak is much reduced, implying that the DP4 product may be preferentially utilized by the enzyme as an acceptor for the glucose moiety (Fig. 20a). In addition, at the slower flow rate of 0.1 ml/min, more than one glucose unit is coupled to a subset of the primer molecules at each passage (Fig. 20b).

11.2.3 Site-Specific Immobilization Compared to the two aforementioned methods, site-specific immobilization can offer both highly specific orientation and reliable stability of the immobilized enzyme (Fig. 17b). The active sites of the immobilized enzymes face away from the support surface and a consequent higher activity can be obtained (Butterfield et al. 2001). A typical site-specific process utilizes the biotin–avidin interaction to immobilize enzymes. Firstly, a single biotin moiety is incorporated onto the appropriate place (opposite the active site) of the enzymes. The biotin is recognized by avidin, which is immobilized onto the membrane surface, and the enzymes are immobilized to the membrane in a specific orientation (Fig. 21). As an advanced technique, site-specific immobilization provides a very promising way to realize simultaneously high stability and activity. Unfortunately, site-specific

Fig. 21. Site-specific immobilization of enzyme by the biotin–avidin interaction. β-Gal βGalactosidase


functionalized enzymes are not easy or readily produced and are usually very expensive. Thus, to our knowledge, no study has been carried out on PPMM with this technique. However, it is predictable that with the current rapid advances in biology this technique will receive more and more consideration in the near future.

11.3 Conclusions The surface of a material is the phase boundary that resides between the bulk material and the outer environment. The performance of materials relies largely upon the properties of the boundaries in many applications. For biological/biomedical applications, polypropylene membranes often suffer from grievous problems such as hydrophobicity, being chemical inert, and lacking functional groups. These disadvantages cause nonspecific protein adsorption and make PPMMs incapable of enzyme immobilization. However, as we can see, graft polymerization of some polymers carrying various functional groups (such as −OH, −NH3 , and −OOH) to the membrane surface is an effective way to conquer these shortcomings. We can choose different modifiers to confer upon the membrane different specialties for diverse applications. Hydrophilic polymers, such as poly(ethylene glycol), polyHEMA, and PAA) are usually used to reduce nonspecific protein adsorption because that protein adsorption is mediated mainly by hydrophobic interactions between the protein and the membrane surface. Nevertheless, it is more intricate for enzyme immobilization due to the complexity of the enzyme itself. Lipase has an affinity for hydrophobic surfaces; however, enzymes that catalyze redox reactions benefit from semiconducting materials. Modifiers must be adapted to a given enzyme in order to maintain its activity. Acknowledgements. Financial support from the National Natural Science Foundation of China (Grant no. 20474054 and 20074033) and the National Basic Research Program of China (Grant no. 2003CB15705) are gratefully acknowledged. The authors thank Dr. Zhen-Mei Liu, Dr. Hong-Tao Deng, and Dr. Rui-Qiang Kou very much for their contribution.

References Aleixo JAG, Swaminathan B, Minnich SA, Wallshein VA (1985) Enzyme immunoassay: binding of Salmonella antigens to activated microtiter plates. J Immunoassay 6:391–407 Annis DA, Jacobsen EN (1999) Polymer-supported chiral co(salen) complexes: synthetic applications and mechanistic investigations in the hydrolytic kinetic resolution of terminal epoxides. J Am Chem Sci 121:4147–4154

296

Q. Yang et al.

Becker M, Provart N, Lehmann I, Ulbricht M, Hicke H (2002) Polymerization of glucans by enzymatically active membranes. Biotechnol Progr 18:964–968 Bossi A, Piletsky SA, Turner APF, Righetti PG (2002) Repartition effect of aromatic polyaniline coatings on the separation of bioactive peptides in capillary electrophoresis. Electrophoresis 23:203–208 Butterfield DA, Lee J, Ganapathi S, Bhattacharyya D (1994) Biofunctional membranes IV. Active site structure and stability of an immobilized enzyme, papain, on modified polysulfone membranes studies by electron paramagnetic resonance and kinetics. J Membr Sci 91:41–52 Butterfield DA, Bhattacharyya D, Daunert S, Bachas L (2001) Catalytic biofunctional membranes containing site-specifically immobilized enzyme arrays: a review. J Membr Sci 181:29–37 Chen Y, Kang ET, Neoh KG, Tan KL (2000a) Covalent immobilization of invertase onto the surface-modified polyaniline from graft copolymerization with acrylic acid. Eur Polym J 36:2095–2103 Chen Y, Kang ET, Neoh KG, Wang P, Tan KL (2000b) Surface modification of polyaniline film by grafting of poly(ethylene glycol) for reduction in protein adsorption and platelet adhesion. Synth Met 110:47–55 Cleveland PH, Wicham GM, Goldbaum MH, Ryan AF, Worthen DM (1981) Rapid and efficient immobilization of soluble and small particulate antigens for solid phase immunoassays. J Immunoassay 2:117–136 Deng HT, Xu ZK, Huang XJ, Wu J, Seta P (2004a) Adsorption and activity of Candida rugosa lipase on polypropylene hollow fiber membrane modified with phospholipid analogous. Langmuir 20:10168–10173 Deng HT, Xu ZK, Liu ZM, Wu J, Ye P (2004b) Adsorption immobilization of Candida rugosa lipases on polypropylene hollow fiber microfiltration membranes modified by hydrophobic polypeptides. Enzyme Microb Technol 35:437–443 Deng HT, Xu ZK, Wu J, Ye P, Liu ZM, Seta P (2004c) A comparative study on lipase immobilized polypropylene microfiltration membranes modified by sugar-containing polymer and polypeptide. J Mol Catal B Enzym 28:95–100 Deppisch R, Storr M, Buck R, Gohl H (1998) Blood material interactions at the surfaces of membranes in medical applications. Sep Purif Technol 14:241–254 Ganapathi S, Butterfield DA, Bhattacharyya D (1998) Kinetics and active fraction determination of a protease enzyme immobilized on functionalized membranes: mathematical modeling and experimental results. Biotechnol Progr 14:865–873 Garber SB, Kingsbury JS, Gray BL, Hoveyda AH (2000) Efficient and recyclable monomeric and dendritic ru-based metathesis catalysts. J Am Chem Sci 122:8168-8179 Gekas VC (1986) Artificial membranes as carriers for the immobilization of biocatalysts. Enzyme Microb Technol 8:450–460 Guell C, Czekaj P, Davis RH (1999) Microfiltration of protein mixtures and the effects of yeast on membrane fouling. J Membr Sci 155:113–122 Gupta B, Anjum N (2003) Plasma and radiation-induced graft modification of polymers for biomedical applications. Adv Polym Sci 162:35–61 Ikada Y (1994) Surface modification of polymers for medical applications. Biomaterials 15:725–726 Kang JS, Shim JK, Huh H, Lee YM (2001) Colloidal adsorption of bovine serum albumin on porous polypropylene-g-poly(2-hydroxyethyl methacrylate) membrane. Langmuir 17:4352–4359 Kasemo B (2002) Biological surface science. Surf Sci 500:656–677

11 Modulation of Adsorption and Activity of Protein/Enzyme on PPMM Surface 297 Kiaei D, Hoffman AS, Horbett TA (1995) Radio-frequency gas discharge (RFGD) fluorination of polymers: protein and cell interactions at RFGD-fluorinated interfaces. Radiat Phys Chem 46:191–197 Kita H, Inada T, Tanaka K, Okamoto K (1994) Effect of photocrosslinking on permeability and permselectivity of gases through benzophenone-containing polyimide. J Membrane Sci 87:139–147 Klee D, Hocker H (1999) Polymers for biomedical applications: improvement of the interface compatibility. Adv Polym Sci 149:1–57 Kou RQ, Xu ZK, Deng HT, Liu ZM, Seta P, Xu YY (2003) Surface modification of microporous polypropylene membranes by plasma-induced graft polymerization of γ -allyl glucoside. Langmuir 19:6869–6875 Kuberkar VT, Davis RH (2000) Modeling of fouling reduction by secondary membranes. J Membrane Sci 168:243–258 Lee JH, Park JW, Lee HB (1991) Cell adhesion and growth on polymer surfaces with hydroxyl groups prepared by water vapour plasma treatment. Biomaterials 12:443–448 Liu ZM, Xu ZK, Wang JQ, Yang Q, Wu J, Seta P (2003) Surface modification of microporous polypropylene membranes by the grafting of poly(γ -stearyl-l-glutamate). Eur Polym J 39:2291–2299 Liu ZM, Xu ZK, Wang JQ, Wu J, Fu JJ (2004) Surface modification of polypropylene microfiltration membranes by graft polymerization of N-vinyl-2-pyrrolidone. Eur Polym J 40:2077–2087 Lundstrom I, Elwing H (1990) Simple kinetic models for protein exchange reactions on solid surfaces. J Colloid Interface Sci 136:68–80 Ma H, Bowman CN, Davis RH (2000a) Membrane fouling reduction by backpulsing and surface modification. J Membr Sci 173:191–200 Ma H, Davis RH, Bowman CN (2000b) A novel sequential photoinduced living graft polymerization. Macromolecules 33:331–335 Malinauskas A (1999) Electrocatalysis at conducting polymers. Synth Met 107:75–83 Mueller J, Davis RH (1996) Protein fouling of surface-modified polymeric microfiltration membranes. J Membr Sci 116:47–60 Oehr C, Muller M, Elkn B, Vohrer U (1999) Plasma grafting – a method to obtain monofunctional surfaces. Surf Coat Tech 116:25–35 Piletsky S, Piletska E, Bossi A, Turner N, Turner A (2003) Surface functionalization of porous polypropylene membranes with polyaniline for protein immobilization. Biotechnol Bioeng 82:86–92 Poche DS, Daly WH, Russo PS (1995) Synthesis and some solution properties of poly(γ -stearyl-L-glutamate). Macromolecules 28:6745–6763 Ratner BD, Hoffman AS, Hanson SR, Harker LA, Whiffen JD (1979) Blood-compatibilitywater-content relationships for radiation-grafted hydrogels. J Polym Sci Polym Symp 66:363–375 Ratner BD (1995) Surface modification of polymers: chemical, biological and surface analytical challenges. Biosens Bioelectron 10:797–804 Richey T, Iwata H, Oowaki H, Uchida E, Matsuda S, Ikada Y(2000) Surface modification of polyethylene balloon catheters for local drug delivery. Biomaterials 21:1057–1065 Sano S, Kato K, Ikada Y (1993) Introduction of functional groups onto the surface of polyethylene for protein immobilization. Biomaterials 14:817–822 Senogles E, Thomas R (1975) Polymerization kinetics of N-vinyl pyrrolidone. J Polym Sci Symp 49:203–210 Sonderquist ME, Walton AG (1980) Structural change in proteins adsorbed on polymer surfaces. J Colloid Interface Sci 75:386–397

298

Q. Yang et al.

Steen ML, Jordan AC, Fisher ER (2002) Hydrophilic modification of polymeric membranes by low temperature H2 O plasma treatment. J Membr Sci 204:341–357 Taniguchi M, Pieracci J, Samsonoff WA, Belfort G (2003) UV-assisted graft polymerization of synthetic membranes: mechanistic studies. Chem Mater 15:3805–3812 Tanioka A, Yokoyama Y, Miyasaka, K (1998) Preparation and properties of enzymeimmobilized porous polypropylene films. J Colloid Interface Sci 200:185–187 Thom V, Altankov G, Groth T, Jankova K, Jonsson G, Ulbricht M (2000) Optimizing cellsurface interactions by photografting of poly(ethylene glycol). Langmuir 16:2756–2765 Tie Y, Calonder C, Van Tassel PR (2003) Protein adsorption: kinetics and history dependence. J Colloid Interface Sci 268:1–11 Ulbricht M, Richau K, Kamusewitz H (1998) Chemically and morphologically defined ultrafiltration membrane surfaces prepared by heterogeneous photo-initiated graft polymerization. Colloid Surf A 138:353–366 Wang Y, Kim JH, Choo KH, Lee YS, Lee CH (2000) Hydrophilic modification of polypropylene microfiltration membranes by ozone-induced graft polymerization. J Membr Sci 169:269–276 Wavhal DS, Fisher ER (2003) Membrane surface modification by plasma-induced polymerization of acrylamide for improved surface properties and reduced protein fouling. Langmuir 19:79–85 Wetzels GMR, Koole LH (1999) Photoimmobilisation of poly(N-vinylpyrrolidinone) as a means to improve haemocompatibility of polyurethane biomaterials. Biomaterials 20:1879–1887 Woltinger J, Bommarius AS, Drauz K, Wandrey C (2001) The chemzyme membrane reactor in the fine chemicals industry. Org Proc Res Dev 5:241–248 Xu ZK, Kou RQ, Liu ZM, Nie FQ, Xu YY (2003) Incorporating r-allyl glucoside into polyacrylonitrile by water-phase precipitation copolymerization to reduce protein adsorption and cell adhesion. Macromolecules 36:2441–2447 Xu ZK, Dai QW, Wu J, Huang XJ, Yang Q (2004) Covalent attachment of phospholipid analogous polymers to modify a polymeric membrane surface: a novel approach. Langmuir 20:1481–1488 Yang Q, Xu ZK, Dai ZW, Wang JL, Ulbricht M (2005) Surface modification of polypropylene microporous membranes with a novel glycopolymer. Chem Mater 17:3050–3058 Zhuang P, Butterfield DA (1992) Structural and enzymatic characterizations of papain immobilized onto vinyl alcohol/vinyl butyral copolymer membrane. J Membr Sci 66:247–257 Zhuang P, Butterfield DA (1993) Optimization of covalently coupling enzymes to polymeric membranes: EPR studies of papain. J Appl Polym Sci 47:1329–1342

12

Nonbiofouling Surfaces Generated from Phosphorylcholine-Bearing Polymers Yasuhiko Iwasaki, Nobuo Nakabayashi, Kazuhiko Ishihara

Abstract. The preparation and characterization of nonbiofouling surfaces generated from phosphorylcholine (PC)-bearing polymers are described. It is proposed that PC groups are an optimum surface with which to create biointerfaces because the surfaces have similarities with biomembranes. Nonspecific protein adsorption is generally the first process to occur when polymers come into contact with the vital environment, and this induces unfavorable bioreactions. Due to the properties of PC surfaces, such as high hydrophilicity, water structure, and zero ζ-potential, protein adsorption is effectively reduced on the PC-bearing surface. This surface property may be important for biomedical applications. The well-defined design of PC surfaces is also introduced in this chapter. These surfaces may be microor nanofabricated for, for example, medical devices and sensors. Control of cell–material interactions is effective on PC-bearing surfaces due to the negligible nonspecific interactions.

12.1 Introduction Protein adsorption is the first phenomenon that occurs when synthetic materials come into contact with a living organism. The uncontrolled protein adsorption functions as a trigger for foreign body reactions to materials from a host. For biomedical applications, control of protein adsorption becomes quite important in the preparation of synthetic materials. Many concepts have been proposed for non-protein-fouling surfaces using physicochemical, biochemical, and biological approaches. One of the most robust approaches is phosphorylcholine immobilization as a mimicker of a biomembrane (Iwasaki and Ishihara 2005). A well-known model of the Yasuhiko Iwasaki, Nobuo Nakabayashi: Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, 2-3-10 Kanda-surugadai, Chiyoda-ku, Tokyo 101-0062, Japan, E-mail: [email protected] Kazuhiko Ishihara: Department of Materials Engineering, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan Principles and Practice Proteins at Solid–Liquid Interfaces Philippe Déjardin (Ed.) © Springer-Verlag Berlin Heidelberg 2006

300

Y. Iwasaki et al.

Fig. 1. Fluid-mosaic model of biomembrane

structure of a biomembrane is the fluid-mosaic model (Fig. 1; Singer and Nicolson 1972). According to this model, amphiphilic phospholipids are arranged in a bilayer structure and proteins are located in or upon it. In all cells for which lipid compositional asymmetry has been described, negatively charged phospholipids such as phosphatidylserine are found predominantly on the inner, cytoplasmic side of the membrane, whereas the neutral, zwitterionic phosphorylcholine lipids such as phosphatidylcholines are located in the outer leaflet. The phosphorylcholine surface provides an inert surface for biological reactions of proteins and glycoproteins to occur smoothly on the membrane. This fact provides very significant information in the development of nonfouling polymer surfaces. In this chapter, a variety of methodologies for making biomimetic phosphorylcholine-bearing surfaces, including well-defined surface preparation (i. e., self-assembled monolayers, SAMs) and polymer brushes, and the surface characteristics of the surfaces are introduced. In addition, an explanation is given as to how securing protein adsorption onto the surface enables the control of cell–material interactions.

12.2 Forces Involved in Protein Adsorption The interactions that are involved in protein adsorption to solid surfaces (Lee et al. 2001) are known to include van der Waals forces, electrostatic “double-layer” forces, solvation forces, and entropic forces (Israelachivili and McGuiggan 1988; Oscarsson 1997). The extent of protein adsorption is

12 Nonbiofouling Surfaces Generated from Phosphorylcholine-Bearing Polymers 301

Fig. 2. Types of attractive and repulsive forces

determined by competition between attractive interactions and nonspecific repulsion (Fig. 2). Among these forces, van der Waals and electrostatic forces are fundamental, while others are introduced by structural or conformational changes. The electrostatic attractions are not likely to play a significant role since most biomaterial surfaces are made of electrically neutral materials. In addition, the Debye length in physiological fluids is usually less than 10 Å. Specific interactions are also expected to be absent, since biomaterial surfaces do not generally possess structures that can be recognized by molecules with specific tertiary structures, such as binding sites of antibodies and enzymes. In general, specific interactions are highly attractive. For example, the free energy of the binding of dinitrophenol by its specific antibodies ranges from –35 kJ/mol to –72 kJ/mol (Bongard 1988). In the absence of such interactions, however, van der Waals forces and hydrophobic interactions are most important. Between hydrophobic surfaces, hydrophobic interactions generally operate over greater distances than does the van der Waals force (Pashley et al. 1985). As mentioned earlier, of the repulsive interactions, electrostatic repulsion is not significant in physiological conditions such as in body fluids or blood. Repulsive hydration forces arise whenever water molecules bind to the surface containing hydrophilic groups (Israelachvili 1985). Steric repulsion can provide long-range repulsion; the long-range aspect is provided by the thickness of the grafted layer of hydrophilic molecules. Thus, it is mostly the long-range steric repulsion that effectively overcomes the attractive interactions in the physiological milieu. If the grafted layer is rather thin (i. e., if the range of the steric repulsion is rather short), the hydration repulsion becomes a significant portion of the overall repulsion. In that case it may not be easy to distinguish between the osmotic factor of the steric repulsion and the hydration forces (Israelachvili 1985).

302

Y. Iwasaki et al.

12.3 Design of Phosphorylcholine-Bearing Surfaces What is the most appropriate surface for controlling interactions with proteins? Although this question has been considered for a long time in relation to the creation of biomedical materials, we have not succeeded in obtaining the expected surface because most surface design is approached from a physicochemical perspective. In contrast, biomembrane surfaces may possess the most preferable characteristics for controlling complex biological interactions. Recently, use of biomembrane structures has been adopted for preparing the surfaces of biomaterials. In an attempt to understand biocompatibility, Nakabayashi and coworkers (Kadoma et al. 1978) first synthesized a methacrylate monomer with a phosphorylcholine group, 2-methacryloyloxyethyl phosphorylcholine (MPC), to obtain new medical polymer materials to mimic biomembrane surfaces (Fig. 3). In 1982, Nakaya et al. also succeeded in synthesizing MPC (Umeda et al. 1982). However, at those times, the degree of purity and yield of MPC was insufficient to allow evaluation of their functions. Ishihara et al. (1990) then developed a new synthetic route and succeeded in producing MPC with good yield as a white powder by recrystallization. MPC, which contains the polymerizable methacrylate group, can readily be copolymerized, enabling the design of numerous polymers with a wide range of molecular architectures including random (Ishihara et al. 1990; Ueda et al. 1992), block (Kojima et al. 1991; Ma et al. 2003a, b; Li et al. 2003), graft (Ishihara et al. 1994b; Iwasaki and Akiyoshi 2004), charged (Ishihara et al. 1994a; Ito et al. 2003), and end-functional polymers (Ishihara et al. 1994b). The homopolymer of MPC [poly(MPC)] is soluble in water. The solubility of MPC polymers can easily be altered by changing the structure and fraction of the comonomers.

Fig. 3. Chemical structure of 2-methacryloyloxyethyl phosphorylcholine (MPC)


12.4 Mechanism of Resistance to Protein Adsorption on the MPC Polymer Surface To reduce protein adsorption on polymer surfaces, surface modifications that induce repulsive interactions or reduce attractive interactions have been conducted. It has been known for quite some time that in general, as a surface becomes more hydrophobic, the extent of protein adsorption increases (Elwing et al. 1987; Golander and Pitt 1990; Golander et al. 1990; Lee and Lee 1993). The unusually strong attraction between proteins and hydrophobic surfaces in water has been described by a hydrophobic interaction, which arises primarily from the structural rearrangement of water molecules in the overlapping solvation zones as proteins adsorb to the surface. Thus, the hydrophobic interaction is primarily an entropic phenomenon without any specific associated bond. The orientation of water molecules adjacent to hydrophobic surfaces is entropically unfavorable. Upon adsorption of proteins to hydrophobic surfaces, the entropically unfavorable water is released in bulk, thereby reducing the total free energy. The hydrophobic interaction then results in the strong attraction of proteins toward the hydrophobic surface. To weaken the hydrophobic interaction, hydrophilic polymers [e. g., poly(2-hydroxyethyl methacrylate) (HEMA), poly(acryl amide), poly(vinylpyrrolidone)] have been applied for surface modification. While these hydrophilic polymers could relatively reduce protein adsorption, their use precludes the accurate control of protein adsorption. Recently, it has been reported that the structure of water absorbed in the polymer materials influences protein adsorption on their surfaces. When protein molecules adsorb to a polymer surface, water molecules between the protein and the surface must be displaced, as shown in Fig. 4 (Lu et al. 1991). A repulsive solvation interaction arises whenever water molecules are associated with a surface that contains hydrophilic groups. Its strength depends on the energy necessary to disrupt the ordered water structure and ultimately dehydrate the surface. Even on a hydrophilic surface, the water structure has a major effect on protein adsorption and subsequent conformational change. Tsuruta (1996) reported that the random networks of water molecules on the material surface are very important in explaining protein adsorption. Protein adsorption processes are considered to start with protein trapping by the random networks of water molecules on the material surface. The material surface, which cannot undergo hydrogen bonding with water, will then reduce protein adsorption. Table 1 lists the free water concentration in the hydrated polymer membrane with a 0.36 water fraction determined

304

Y. Iwasaki et al.

Fig. 4. Schematic description of protein adsorption on the polymer surface. Water molecules are moved from contact sites between amino acid residues and the polymer surface. A change in the conformation of the adsorbed protein then occurs and the protein binds to the surface tightly. ∆F Free energy, ∆H enthalpy, T absolute temperature, ∆S entropy Table 1. Characteristics of hydration state of polymer. Heq = (weight of water in the polymer membrane)/(weight of polymer membrane saturated with water) at 25 ◦ C, Poly(HEMA) poly(2-hydroxyethyl methacrylate), PMB poly(MPC-co-n-butyl methacrylate), Number after PMB is molar percent of MPC in the coplymer Poly(HEMA) Heq 0.40 Free water fraction at Heq 0.34 At H = 0.36 0.28

10

PMB 30

0.23 0.84 0.25 0.84 – 0.69

with differential scanning calorimetry (Ishihara et al. 1998). The fraction of free water (not bound water) in the MPC polymer was 0.65, which was found to be significantly higher than that in poly(HEMA), which was 0.28. In addition, the structure and hydrogen bonding of water in the vicinity of poly(MPC-co-butyl methacrylate) (PMB) were analyzed in their aqueous solutions and thin films with contours of O-H stretching of the Raman and attenuated total reflection infrared (ATR-IR) spectra, respectively (Kitano et al. 2000, 2003). Figure 5 shows the O-H stretching Raman band of water in various polymer solutions recorded in the region between 2500 and 4000/cm by using the polarization method. For the polarization geometries X(ZZ)Y (parallel position, I// ) and X(ZX)Y (perpendicular position, I⊥ ), a polarizer plate was rotated exactly 90 ◦ C in front of the slit, where X and Y are the directions of the laser beam and observation, respectively. A depolarization ratio, r, is an indicator of symmetry of the vibration mode, and is expressed by r = I// / I⊥ , where I// and I⊥ are the intensities of the spectra observed with the polarizer oriented parallel and perpendicular to the incident laser beam, respectively. The component of the O–H


Fig. 5. Raman shifts of O–H stretching region. a I// and I⊥ spectra of pure water at 25 ◦ C, where I// and I⊥ are the intensities of the spectra observed with the polarizer oriented parallel and perpendicular to the incident laser beam, respectively. The I⊥ /ρ spectrum (where ρ is the depolarization ratio) is shown with a dotted line. b The collective band of water. Reprinted with permission from Kitano et al. (2000), copyright (2000) American Chemical Society

stretching band of water centered at 3250/cm was highly polarized and diminished in the spectra at the perpendicular position. The polarized O–H stretching band of water, which is called a collective band, is ascribed to an H2 O molecule executing u1 vibrations all in phase with each other but with a vibrational amplitude varying from molecule to molecule in water clusters, which are strongly hydrogen-bonded. Theoretical calculations of a random network model, which is characterized by fluctuating defects in water–water hydrogen bonds in a distorted tetrahedral network, support the interpretation.

306

Y. Iwasaki et al.

To clarify the effect of polymers on the structure of water, the intensity of the collective band (Ic ) observed at around 3250/cm was separated from the spectra using Eq. 1 (Fig. 5): Ic = I// − I⊥ / rO–H

(1)

Since the intensities of Raman spectra are not absolute, the area of Ic was normalized as Eq. 2: C = ∫ Ic (w)dw/ ∫ I⊥ (w)dw = Cx / Cw

(2)

where w is the Raman shift in cm−1 . Figure 6 shows the effect of the molecular weight (Mw )of different kinds of polymers on the relative intensities of the collective bands (C) at a constant molar fraction of monomer units (px = 0.01 or 0.05). In the case of aqueous sodium poly(styrenesulfonate) solution, a very strong fluorescence background made the Raman spectroscopic measurements impossible. The values of C for other polymer solutions were almost constant in the region of relatively low Mw , and decreased with an increase in Mw in the region where the Mw values were larger than some critical values. In general, the whole concentration region of polymer solutions can be divided into three parts: dilute, semidilute, and concentrated. The macroscopic properties of polymer solutions in the different concentration regions are

Fig. 6. Plots of values of the relative intensities of the collective bands (C) for various aqueous polymer solutions at molar fraction (px ) = 0.01 and 25 ◦ C versus molecular weight (Mw ). Sodium poly(styrenesulfonate) (open triangles), poly-l-lysine hydrobromide (open circles), poly(2-methacryloyloxyethyl phosphorylcholine) [poly(MPC); filled squares], poly(N-vinylpyrrolidone) (filled triangles), and poly(ethylene glycol) (filled circles; px = 0.05). Reprinted with permission from Kitano et al. (2000), copyright (2000) American Chemical Society


significantly altered. The figure seems to show that the structure of water in a dilute polymer solution and in a semidilute polymer solution are different, as discussed below. The C value for the aqueous solution of poly(MPC)s was very close to that for pure water, as shown in Fig. 6, which is in contrast with the smaller C value in the aqueous solution of ordinary polyelectrolytes. A similar tendency was also observed on hydrated thin polymer films. These results suggest that the PMB does not significantly disturb the hydrogen bonding between the water molecules in either the aqueous solution or the thin film systems. The equilibrium amount of the proteins bovine serum albumin (BSA) and bovine plasma fibrinogen (BPF) adsorbed on the polymer surface was measured and represented with the free water fraction in the hydrated polymers, as shown in Fig. 7 (Ishihara 2000). The amounts of both proteins adsorbed on poly(HEMA), poly(acryl amide (AAm)-co-n-butyl methacrylate) (BMA), and poly(N-vinylpyrrolidone (VPy)-co-BMA) were larger than were those on poly(MPC-co-dodecyl methacrylate) (PMD) and PMB. It was reported that the theoretical amount of BSA and BPF adsorbed on the surface in a monolayer state were 0.9 and 1.7 µg/cm, respectively. On the surface of the MPC polymers, the amount of adsorbed proteins was less than these theoretical values.

Fig. 7. Relationship between free water fraction in hydrated polymer membrane and amount of proteins adsorbed on the polymer. Concentration of bovine serum albumin ([BSA]) in phosphate-buffered saline (PBS) = 0.45 g/dl (squares); concentration of bovine plasma fibrinogen ([BPF]) = 0.30 g/dl (circles). Poly(HEMA) Poly(2-hydroxyethyl methacrylate), Poly(AAm-co-BMA) poly(acryl amide-co-n-butyl methacrylate), Poly(VPy-c-BMA) poly(N-vinylpyrrolidone-co-butyl methacrylate), PMD poly(MPC-co-dodecyl methacrylate), PMB poly(MPC-co-butyl methacrylate)

308

Y. Iwasaki et al.

Electrostatic interactions between proteins and solid surfaces have also been discussed. Almost all proteins in plasma are negatively charged under physiological conditions because their isoelectric points are below pH 7.4. Therefore, plasma proteins adsorbed favorably onto positively charged surfaces. Holmlin and coworkers (2001) compared surface ability to resist nonspecific protein adsorption on zwitterionic SAMs, referring both to SAMs formed from a 1:1 mixture of positively and negatively charged thiols and thiols combined in a positively and negatively charged moiety in the same molecule. As shown in Fig. 8, the amount of adsorbed protein on zwitterionic SAMs was much less than that on positively charged or negatively charged SAMs. Moreover, single-component SAMs formed from thiols ter-

Fig. 8. Plots of the change in response unit (∆RU) or irreversible adsorption of fibrinogen (a) and lysozyme (b) to different self-assembled monolayers (SAMs) as a function of the ionic strength of the buffer dissolving the protein. The buffer was 4.4 mM phosphate (pH 7.4, ionic strength 10 mM); the ionic strength was adjusted by dissolving NaCl in the appropriate concentrations. The symbols corresponding to the different functional groups presented at the SAM–buffer interface are defined above the plots. Reprinted with permission from Holmlin et al. (2001), copyright (2001) American Chemical Society


minating in groups combining a positively charged moiety and a negatively charged moiety were capable of resisting the adsorption of proteins. Poly(ethylene oxide) (PEO) is one of the most widely used hydrophilic polymers for surface modification of biomaterials (Harris 1992). The high water solubility of PEO is the result of a good structural fit between water molecules and the polymer. Steric repulsion and molecular flexibility are believed to be the dominant factors for reduction protein adsorption on PEO-immobilized surfaces. However, surface modification with few highly dense ethylene oxide oligomers is also effective in reducing protein adsorption (Johnston et al. 2005; Wu et al. 2000). The length of such oligomers is on the order of 1 nm. These results suggest that steric repulsion is not the dominating factor in the prevention of protein adsorption by surface modification. A similar behavior can be seen on a phosphorylcholine surface. In previous literature on MPC copolymer systems, chain mobility was thought to be the dominant factor in resisting protein adsorption onto a surface. Conversely, high-density MPC polymer brushes also reduced protein adsorption. The molecular mobility of a polymer brush, which is defined by the hysteresis between the contact angles of advancing and receding water, was much lower than that of the MPC copolymer. The surface characteristics of a phosphorylcholine polymer surface for the reduction of protein adsorption are summarized in Fig. 9. The phos-

Fig. 9. Possible factors for the nonfouling property of phosphorylcholine-bearing surfaces. ζ-potential Charge potential

310

Y. Iwasaki et al.

phorylcholine group is very hydrophilic, which means it is unfavorable for attracting hydrophobic interactions with proteins. Moreover, the phosphorylcholine polymer does not disintegrate the water structure around the polymer. This is a unique property of phosphorylcholine polymers in comparison with conventional hydrophilic polymers. In addition, electrostatic interactions between a phosphorylcholine polymer surface and proteins are weak because the charge potential (ζ-potential) of phosphorylcholine is neutral (Ishihara et al. 1994a). Therefore, phosphorylcholine polymer has several factors for reducing nonspecific protein adsorption. This polymer is one of the best polymer materials for making a nonbiofouling surface.

12.5 Fundamental Interactions Between MPC Polymers and Proteins Protein adsorption on material surfaces causes serious biological reactions such as, for example, thrombus formation, immune response, and complement activation, capsulation (Brash and Horbett 1987; Horbett and Brash 1995). To understand the blood compatibility of surfaces, it is necessary not only to determine the amount of adsorbed protein but also the species

Fig. 10. Amount of protein adsorbed on polymer surfaces from human plasma. Poly(BMA) poly(butyl methacrylate), Poly(HEMA) poly(2-hydroxyethyl methacrylate), PMB poly(MPC-co-n-butyl methacrylate). Number after PMB is molar percent of MPC in the coplymer


of the protein. Therefore, the effects of MPC units on protein adsorption were investigated. Figure 10 shows the amount of protein adsorbed on PMB, poly(HEMA), and poly(butyl methacrylate) (BMA) after contact with plasma for 60 min (Ishihara et al. 1992). On poly(BMA), many more proteins were adsorbed than on either poly(HEMA) or PMB. The amount of proteins adsorbed on PMB decreased with increases in the ratio of the MPC composition of the polymer. The species and distribution of the protein adsorbed on PMB were also determined by gold-colloid and radiolabeled immunoassay (Ishihara et al. 1991b). From these experiments, it was clarified that the PMB could reduce plasma protein adsorption nonspecifically. Thrombus formation on conventional polymeric materials occurred through the multilayers of plasma proteins denatured by contact with the surfaces. The secondary structures of BSA and BPF adsorbed onto the PMB were evaluated by circular dichroism (CD) spectroscopy (Ishihara et al. 1991a, 1998). Figure 11 shows the CD spectra of BSA in phosphate-buffered saline (PBS) and that adsorbed on the polymer surface. For BSA in PBS, the mean molecular residual ellipticity had a large negative value at 222 nm. The CD spectrum of BSA adsorbed onto PMB was almost the same as that in PBS. The negative ellipticity at 222 nm of BSA adsorbed on the MPC polymers increased as the ratio of MPC decreased, then became almost zero for BSA adsorbed on poly(HEMA). The authors found the same tendency in the case of BPF. Calculation of the α-helix contents of BSA and BPF revealed that the PMB could effectively suppress the conformational change of proteins even when the proteins were adsorbed on the surface (Ishihara et al. 1991, 1998).

Fig. 11. Circular dichroism spectra of BSA in PBS and of that adsorbed on polymer surfaces. A poly(HEMA), B PMB10, C PMB30, D BSA/PBS. Mol. Ellip. Molar ellipticity

312

Y. Iwasaki et al.

In contrast, the α-helix content of both proteins adsorbed on poly(HEMA) decreased significantly. The protein-adsorption-resistant properties of the MPC polymer have also been determined by other researchers (Campbell et al. 1994; Chang et al. 1998; Sugiyama et al. 1997).

12.6 Recent Designs of Nonfouling Phosphorylcholine Surfaces with Well-Defined Structures MPC is one of the best monomers with which to produce biomimetic surfaces because it can be applied to a wide variety of surface modifications, as shown in Fig. 12. These methods have been applied to improve the bio/blood compatibility of numerous biomedical devices (Iwasaki and Ishihara 2005). To better understand protein–material and cell–material interactions at the submolecular level, well-defined biomimetic surfaces have recently been produced (Fig. 13). Alkanethiols terminating in various functional groups have been used to study the physical-organic chemistry of the adsorption of proteins to synthetic surfaces (Kohler at al. 1996; Marra et al. 1997; Tegoulia and Cooper 2000; Wang et al. 2000). It is clear that alkanethiols fixed on a gold surface create a SAM. Tegoulia and Cooper (2000) first reported that adhesion of neutrophils was effectively reduced on a SAM surface with phosphorylcholine groups. Chen et al. (2005) synthesized disulfide molecules having phosphorylcholine groups and prepared SAMs on a gold surface.

Fig. 12. Surface modifications of a polymer surface with MPC polymers


Fig. 13. Well-defined designs of phosphorylcholine-bearing surfaces

Marra et al. (1997) synthesized the phospholipid monomer 1-palmitoyl2-[12-(acryloyloxy)dodecanoyl]-sn-glycero-3-phosphorylcholine, prepared as unilamellar vesicles, and fused them onto alkylated glass. Freeradical polymerization was performed in an aqueous solution at 70 ◦ C. It was clear from x-ray photoelectron spectroscopy analysis that the phospholipid assembly had a closely packed monolayer formation. This formation is very stable under static conditions in water and air, as well as in an environment where there is a high shear flow. There are several procedures for the creation of biomembrane-like surfaces with polymerizable phospholipids (Conboy et al. 2003; Feng et al. 2002; Kim et al. 2004). Surface modification using silane coupling has also been demonstrated with reagents bearing phosphorylcholine. The process was classified into two approaches. One is generated from hydroxyl-terminated monolayers on a silicon wafer (Durrani et al. 1986; Hayward et al. 1984). The other is the chemical or physical adsorption of phospholipids or phospholipid derivatives on alkylsilane monolayers (Hayward et al. 1986a, b; Marra et al. 1997; Tegoulia and Cooper 2000). Kohler et al. (1996) prepared a glass surface that reacted with 3-aminopropyl-trimethoxysilane. The carboxylated phosphatidylcholine was grafted with a coupling agent. Matsuda and coworkers (2003a) prepared a phosphorylcholine-endcapped poly(N,N’-dimethylacrylamide) [poly(DMEAA)] and a block co-

314

Y. Iwasaki et al.

oligomer with poly(styrene) with the aid of a photoiniferter-based quasiliving polymerization technique. The oligomer exhibits amphiphilic properties and chemisorbs on a gold surface with hydrophobic anchoring. The surface coated with the oligomers reduced plasma protein adsorption and cell adhesion. The authors also explored a surface design for producing one or two phosphorylcholine groups capped at the terminal end of a graft chain of poly(DMEAA) (Matsuda et al. 2003b). Polymers may have more worth because of their great potential for multiple functionalities. Although preparation of a well-defined surface is normally difficult due to their structural and size distributions, welldefined MPC polymer brushes on silicon wafers have been prepared quite recently by atom transfer radical polymerization (ATRP) (Feng et al. 2004, 2005; Iwata et al. 2004). ATRP is one of the best methods for living radical polymerization because it can be applied to a wide range of monomers. An alternative process, pioneered by Wirth and Tsujii (Huang et al. 1997, 1999; Ejaz et al. 1998, 2002), to prepare well-defined polymer brushes on solid surfaces with ATRP is considerably theoretical and deals with experimental interests in the control of surface properties. Iwata et al. (2004) reported the manipulation of protein adsorption on a thin MPC polymer brush surface. First, they treated silicon wafers with 3-(2-bromoisobutyryl)propyl dimethylchlorosilane (BDCS) to form a monolayer that acts as an initiator for ATRP. Silicon-supported BDCS monolayers were soaked in a methanol/ water solution containing Cu(I)Br, bipyridine, and a sacrificial initiator. After MPC was added to the solution, ATRP was allowed to progress for 18 h. The synthetic scheme for poly(MPC) brushes is shown in Fig. 14a. The Mw and thickness of the poly(MPC) brush layer on the silicon surface increased with increasing polymerization time, as shown in Fig. 14b. The dense polymer brushes were obtained by the “grafting from” system. By selective decomposition of the BDCS monolayer by ultraviolet (UV)-lightirradiation, the poly(MPC) brush region and the sizes were well controlled, resulting in fabricating micropatterned poly(MPC) brushes.

12.7 Control of Cell–Material Interactions on a Phosphorylcholine Polymer Nonfouling Surface Cell adhesion on material surfaces depends strongly on the plasma (serum) protein adsorption onto the surface because there are many kinds of cell adhesive proteins in plasma (serum). It has already been reported that an MPC polymer can reduce the adhesion of a variety of cells such as blood cells (Hasegawa et al. 2002; Ishihara et al. 1992, 1993, 2000; Iwasaki et al. 1997, 2001), fibroblasts (Ishihara et al. 1999; Iwasaki et al. 1999), and


Fig. 14. a Synthetic route of poly(MPC) brushes on silicon wafer via atom transfer radical polymerization. b Thickness of silicon-sputtered poly(MPC) brushes as function of polymerization time

bacteria (Hirota et al. 2005; Patel et al. 2005). By manipulation of the surface distribution of MPC polymers or tag-immobilization to recognize a cell on MPC polymers, the cell attachment on the surface will be well controlled.

12.7.1 Cell Manipulation on a Well-Defined Phosphorylcholine Polymer Brush Adsorption of fluorescein-isothiocyanate-labeled BSA was well controlled on a patterned graft polymer surface, as shown in Fig. 15. On an UVirradiated region with no polymer brush, the fluorescence intensity was significantly high, indicating that a large amount of BSA was adsorbed in this region. Conversely, BSA adsorption was remarkably reduced in the

316

Y. Iwasaki et al.

Fig. 15. Fluorescence micrographs of fluorescein isothiocyanate-labeled albumin adsorption on patterned poly(MPC) brush surface after contact with 0.45 g/dl FITC-albumin in PBS for 30 min. Reprinted with permission from Iwata et al. (2004), copyright (2004) American Chemical Society

poly(MPC) brush layer. Figure 16 shows the fibronectin adsorption pattern on the pattern surface after contact with the cell culture medium for 60 min. The adsorption pattern was determined by immunoassay. On the polymer brush prepared by polymerization for 10 min, fluorescence caused by the adsorbed fibronectin was observed to be homogeneous. The difference in fibronectin adsorption on the patterned surface was clearly related to the

Fig. 16. Fluorescence micrographs of fibronectin adsorption on a patterned poly(MPC) brush surface after contact with cell culture medium for 30 min. Reprinted with permission from Iwata et al. (2004), copyright (2004) American Chemical Society. L Large, M medium, S small


time of polymerization of the polymer brushes. On the poly(MPC) brush surface, fibronectin adsorption was effectively reduced. The reduction of plasma protein adsorption on MPC polymer surfaces has been reported previously, and the mechanism is considered to be related to the structure of the polymer surface (Ishihara et al. 1998). The study to clarify the difference in surface structure between a PMB coating and the dense poly(MPC) brush prepared in this study on the reduction of protein adsorption is still ongoing. However, it has been shown that poly(MPC) brushes just 5 nm thick prepared by the “grafting from” system can reduce protein adsorption. Although the thickness of the cast film of PMB that we made has been normally controlled on a submicron scale, the poly(MPC)-brush thickness can be controlled on the scale of a few nanometers. This is a great advantage for surface modification to improve the nonfouling properties of micro- or nanodevices. Figure 17 presents fluorescence micrographs of fibroblasts that adhered onto the patterned surface. The effect of surface grafting of poly(MPC) by polymerization for 10 min on fibroblast adhesion was not observed and the cells adhered homogeneously to the surface. By adjusting the time of polymerization, cell adhesion was controlled and fibroblasts adhered to UV-irradiated regions that had no poly(MPC) brushes. This result is coincident with protein adsorption on the surface. Above a poly(MPC)brush thickness of about 5 nm, protein adsorption and cell adhesion was remarkably reduced (i. e., they were able to recognize the thickness of the thin brush). The effect of the grafting density of poly(ethylene glycol) on protein adsorption has been reported by Sofia and coworker (1998). Grafting density may also be important for controlling protein adsorption on poly(MPC)-brush surfaces. The number of adherent cells can also be controlled with a change in the surface area of the UV-irradiated region. The surface area of the pattern affected cell density. The surface areas for adherent cells on smalland medium-sized pattern surfaces were 844 ± 185 µm2 /cell and 1188 ± 240 µm2 /cell, respectively. The large-sized pattern surface had a cell density of 159 ± 34 µm2 /cell, which is significantly higher. Cell communication depends on the area of the surface where the cells are able to adhere. To fabricate a microscale pattern on a solid surface, microcontact printing and soft lithography techniques have been applied. Whitesides and coworkers reported that microscale patterning provides a versatile method for creating novel adhesive substrates that are useful for spatially positioning mammalian cells and controlling their viability, form, and function (Brock et al. 2003; Chen et al. 1997; Singhvi et al. 1994). Armes and coworkers have reported that the molecular architectures of MPC polymers can be easily controlled with ATRP in protic solvents (Li et al. 2003; Lobb et al. 2001; Ma et al. 2002), with the result that a wide vari-

318

Y. Iwasaki et al.

Fig. 17. Fluorescence micrographs of fibroblast adhesion on patterned poly(MPC) brush surface after incubation for 24 h. Concentration of fibroblasts, [Fibroblast] = 5.0×104 cells/ml. Reprinted with permission from Iwata et al. (2004), copyright (2004) American Chemical Society

ety of polymer surface designs will be possible. Surface modification with well-defined MPC polymers would be considered as one of the most robust methods with which to optimize biointerfaces on a molecular scale. Microfabrication with MPC polymers might prove to be important in separations, biosensors, and the development of biomedical materials.

12.7.2 Selective Cell Attachment to a Biomimetic Polymer Surface Through the Recognition of Cell-Surface Tags Synthetic polymers capable of selectively recognizing proteins or cells play important roles in cell separation, biosensors, and the development of biomedical materials (Ratner 1993, 1996). In general, few synthetic poly-


mers can recognize a specific biomolecule or cell in vivo because of the complexity of the environment. In a living cell, carbohydrates on the cell surface contribute to most of the communications between the cell and its environment (Dwek 1996). Therefore, it might be possible to control cell functions with polymers, which can affect carbohydrate interactions with a cell surface. Tagging of carbohydrates on the cell membrane, as well as the polymer design, might be effective for the recognition of a specific cell. The incorporation of unnatural carbohydrates into living cells provides an opportunity to study the specific contributions of sialic acid and its N-acyl side chains to sialic acid-dependent ligand-receptor interactions at a submolecular level. Bertozzi and coworkers have expressed unnatural functional groups (i. e., acetyl, levulinoyl, and azideacetyl groups), on living cell surfaces through the glycosylation of unnatural monosaccharides (Mahal et al. 1997; Saxon and Bertozzi 2000; Yarema and Bertozzi 2001; Yarema et al. 1998). Iwasaki et al. (2005) reported a novel strategy for controlling selective cell attachment and detachment even in nonadhesive cells by using biomimetic polymer surface engineering and cell-surface engineering. To this end, they designed reactive biomimetic MPC polymer surfaces with hydrazide groups, which can react selectively with unnatural ketone-containing carbohydrates as a cell surface tag. Figure 18 shows a schematic representation of selective cell attachment through the recognition of a cell-surface tag. The MPC polymer surface prohibits nonspecific interaction with plasma protein, which often interferes with the specificity of the interaction of materials with a cell membrane. Hydrazide groups on the MPC polymer surface react with ketone groups on the cell surface even in cell culture medium. Surface tags on living cells can be introduced in the presence of N-levulinoylmannosamine (ManLev) by surface engineering (Lee et al. 1999; Lemieux et al. 1999). The chemical structures of MPC polymers to control cell attachment are shown in Fig. 19.

Fig. 18. Schematic representation of specific cell attachment to a nonfouling MPC polymer surface through cell-surface tags

320

Y. Iwasaki et al.

Fig. 19. Chemical structures of PMB, poly(MPC-o-BMA-co-MH) (PMBH), and poly(BMAco-MH) (PBH)

To clarify the possibility of using a polymer surface to recognize specific cells with cell-surface tags, the coculture of ManLev-treated HL-60 cells with native HeLa cells was examined. HeLa cells are typical cancer cells that adhere nonspecifically to conventional polymer surfaces through cellbinding proteins adsorbed on the surface. Figure 20 shows phase-contrast and fluorescence microscopy images of polymer surfaces after the coculture experiment. Cell adhesion on the PMB surface was completely suppressed.

Fig. 20. Phase-contrast and fluorescence micrographs of polymer surfaces after the coculture experiment. White arrows indicate HL-60 cells. Reprinted with permission from Iwasaki et al. (2005), copyright (2005) American Chemical Society


In contrast, both HL-60 and HeLa cells were observed on poly(BMA-coMH) (PBH). HL-60 cells selectively attached to poly(MPC-o-BMA-co-MH) (PMBH), while HeLa cells did not. The adhesion of HeLa to a polymer surface is closely associated with the adsorption of cell-binding proteins on the surface, such as fibronectin, fibrinogen, and vitronectin, because these proteins have an RGD binding site for the integrin of the cell membrane (Hirano et al. 1993). On PMBH, the adhesion of HeLa cells was reduced due to fibronectin resistance. It has been reported previously that MPC polymer can nonspecifically reduce plasma protein adsorption (Ishihara et al. 1998). The nonfouling property of the MPC polymer surface is quite important for achieving specific interactions with a cell. These results indicate that cell-surface recognition by a polymer surface is evident in both single- and two-cell systems. Such recognition on an MPC polymer surface is effective in a multiple-cell system because the MPC polymer surface rejects the adhesion of cells without tags. Moreover, the control of cell attachment can be applied to a wide range of cells because the expression of unnatural carbohydrates on cell surfaces has been observed on numerous types of cells.

12.8 Conclusion The control of protein adsorption on material surfaces is a most fundamental as well as important biomaterial-related research subject. Although surface designs have been performed to control protein–material interactions for many years, the surface structure, which can completely control the interactions, has not yet been optimized. To this end, nonfouling surfaces are required, and biomembrane-mimetic surfaces (e. g., phosphorylcholine polymers) might be one of the most suitable materials. In addition, a welldefined surface would be useful for clearly understanding protein–material interactions at the submolecular level. The study of adsorption is necessary to achieve progress with the next generation of biomaterials and thus secure a material that can control protein.

References Bongard P (1988) Physical Basis of Cell–Cell Adhesion. CRC Press, Boca Raton, FL Brash JL, Horbett TA (eds) (1987) Proteins at Interfaces: Physicochemical and Biochemical Studies. ACS Symposium Series 343. American Chemical Society, Washington, DC Brash JL, Horbett TA (eds) Proteins At Interfaces II: Fundamentals and Applications. American Chemical Society, Washington, DC

322

Y. Iwasaki et al.

Brock A, Chang E, Ho CC, LeDuc P, Jiang X, Whitesides GM, Ingber DE (2003) Geometric determinants of directional cell motility revealed using microcontact printing. Langmuir 19:1611–1617 Campbell EJ, O’Byrne V, Stratford PW, Quirk I, Vick TA, Wiles MC, Yianni YP (1994) Biocompatible surfaces using methacryloylphosphorylcholine laurylmethacrylate copolymer. ASAIO J 40:853–857 Chang PC, Lee SD, Hsiue GH (1998) Heterobifunctional membranes by plasma induced graft polymerization as an artificial organ for penetration keratoprosthesis. J Biomed Mater Res 39:380–389 Chen C, Zheng J, Li L, Jiang S (2005) Strong resistance of phosphorylcholine self-assembled monolayers to protein adsorption: insights into nonfouling properties of zwitterionic materials. J Am Chem Soc 127:14473–14478 Chen CS, Mrksich M, Huang S, Whitesides GM, Ingber DE (1997) Geometric control of cell life and death. Science 276:1425–1428 Conboy JC, Liu S, O’Brien F, Saavedra SS (2003) Planar supported bilayer polymers formed from bis-diene lipids by Langmuir-Blodgett deposition and UV irradiation. Biomacromolecules 4:841–849 Durrani AA, Hayward JA, Chapman D (1986) Biomembranes as models for polymer surfaces. II. The syntheses of reactive species for covalent coupling of phosphorylcholine to polymer surfaces. Biomaterials 7:121–125 Dwek RA (1996) Glycobiology: toward understanding the function of sugars. Chem Rev 96:683–720 Ejaz M, Yamamoto S, Ohno K, Tsujii Y, Fukuda T (1998) Controlled graft polymerization of methyl methacrylate on silicon substrate by the combined use of the Langmuir-Blodgett and atom transfer radical polymerization techniques. Macromolecules 31:5934–5936 Ejaz M, Yamamoto S, Ohno K, Tsujii Y, Fukuda T (2002) Fabrication of patterned highdensity polymer graft surfaces. 1. Amplification of phase-separated morphology of organosilane blend monolayer by surface-initiated atom transfer radical polymerization. Macromolecules 35:1412–1418 Elwing H, Welin A, Askendahl A, Nilsson U, Lundstrom I (1987) A wettability gradient method for studies of macromolecular interaction at the liquid/solid interface. J Colloid Interface Sci 119:203–210 Feng J, Tseng P-Y, Faucher KM, Orban JM, Sun XL, Chaikof EL (2002) Functional reconstitution of thrombomodulin within a substrate-supported membrane-mimetic polymer film. Langmuir 18:9907–9913 Feng W, Brash J, Zhu SP (2004) Atom-transfer radical grafting polymerization of 2methacryloyloxyethyl phosphorylcholine from silicon wafer surfaces. J Polym Sci A Polym Chem 42:2931–2942 Feng W, Zhu S, Ishihara K, Brash JL (2005) Adsorption of fibrinogen and lysozyme on silicon grafted with poly(2-methacryloyloxyethyl phosphorylcholine) via surface-initiated atom transfer radical polymerization. Langmuir 21:5980–5987 Golander CG, Lin YS, Hlady V, Andrade JD (1990) Wetting and plasma-protein adsorption studies using surfaces with a hydrophobicity gradient. Colloids Surf 49:289–302 Golander CG, Pitt WG (1990) Characterization of hydrophobicity gradients prepared by means of radio frequency plasma discharge. Biomaterials 11:32–35 Harris JM (1992) Poly(ethylene glycol) Chemistry. Plenum Press, N.Y Hasegawa T, Iwasaki Y, Ishihara K (2002) Preparation of blood-compatible hollow fibers from a polymer alloy composed of polysulfone and 2-methacryloyloxyethyl phosphorylcholine polymer. J Biomed Mater Res 63:333–341 Hayward JA, Chapman D (1984) Biomembrane surfaces as models for polymer design: the potential for haemocompatibility. Biomaterials 5:135–142

12 Nonbiofouling Surfaces Generated from Phosphorylcholine-Bearing Polymers 323 Hayward JA, Durrani AA, Lu Y, Clayton CR, Chapman D (1986a) Biomembranes as models for polymer surfaces. IV. ESCA analyses of a phosphorylcholine surface covalently bound to hydroxylated substrates. Biomaterials 7:252–258 Hayward JA, Durrani AA, Shelton CJ, Lee DC, Chapman D (1986b) Biomembranes as models for polymer surfaces. III. Characterization of a phosphorylcholine surface covalently bound to glass. Biomaterials 7:126–131 Hirano Y, Okuno M, Hayashi T, Goto K, Nakajima A (1993) Cell-attachment activities of surface immobilized oligopeptides RGD, RGDS, RGDV, RGDT, and YIGSR toward five cell lines. J Biomater Sci Polym Ed 4:235–243 Hirota K, Murakami K, Nemoto K, Miyake Y (2005) Coating of a surface with 2-methacryloyloxyethyl phosphorylcholine (MPC) co-polymer significantly reduces retention of human pathogenic microorganisms. FEMS Microbiol Lett 248:37–45 Holmlin RE, Chen X, Chapman RG, Takayama S, Whitesides GM (2001) Zwitterionic SAMs that resist nonspecific adsorption of protein from aqueous buffer. Langmuir 17:2841– 2850 Horbett TA, Brash JL (1995) Proteins at Interfaces II – Fundamentals and Applications. ACS Symposium Series 602. American Chemical Society, Washington, D.C Huang X, Wirth MJ (1997) Surface-initiated radical polymerization on porous silica. Anal Chem 69:4577–4580 Huang X, Wirth MJ (1999) Surface initiation of living radical polymerization for growth of tethered chains of low polydispersity. Macromolecules 32:1694–1696 Ishihara K (2000) Bioinspired phospholipid polymer biomaterials for making high performance artificial organs. Sci Technol Adv Mater 1:131–138 Ishihara K, Fujita H, Yoneyama T, Iwasaki Y (2000) Antithrombogenic polymer alloy composed of 2-methacryloyloxyethyl phosphorylcholine polymer and segmented polyurethane. J Biomater Sci Polym Ed 11:1183–1195 Ishihara K, Inoue H, Kurita K, Nakabayashi N (1994a) Selective adhesion of platelets on a polyion complex composed of phospholipid polymers containing sulfonate groups and quarternary ammonium groups. J Biomed Mater Res 28:1347–1355 Ishihara K, Ishikawa E, Iwasaki Y, Nakabayashi N (1999) Inhibition of fibroblast cell adhesion on substrate by coating with 2-methacryloyloxyethyl phosphorylcholine polymers. J Biomater Sci Polym Ed 10:1047–61 Ishihara K, Nomura H, Mihara T, Kurita K, Iwasaki Y, Nakabayashi N (1998) Why do phospholipid polymers reduce protein adsorption? J Biomed Mater Res 39:323–330 Ishihara K, Oshida H, Endo Y, Ueda T, Watanabe A, Nakabayashi N (1992) Hemocompatibility of human whole blood on polymers with a phospholipid polar group and its mechanism. J Biomed Mater Res 26:1543–1552 Ishihara K, Oshida H, Endo Y, Watanabe A, Ueda T, Nakabayashi N (1993) Effects of phospholipid adsorption on nonthrombogenicity of polymer with phospholipid polar group. J Biomed Mater Res 27:1309–1314 Ishihara K, Tsuji T, Kurosaki K, Nakabayashi N (1994b) Hemocompatibility on graft copolymers composed of poly(2methacryloyloxyethyl phosphorylcholine) side chain and poly(n-butyl methacrylate) backbone. J Biomed Mater Res 28:225–232 Ishihara K, Ueda T, Nakabayashi N (1990) Preparation of phospholipid polymers and their properties as polymer hydrogel membrane. Polym J 22:355–360 Ishihara K, Ueda T, Saito N, Kurita K, Nakabayashi N (1991a) Suppression of protein adsorption and denaturation on polymer surface with phospholipid polar group. Seitai Zairyou (J J Soc Biomater) 9:243–249 Ishihara K, Ziats NP, Tierney BP, Nakabayashi N, Anderson JM (1991b) Protein adsorption from human plasma is reduced on phospholipid polymers. J Biomed Mater Res 25:1397– 1407

324

Y. Iwasaki et al.

Israelachvili JN (1985) Intermolecular and Surface Forces. Academic Press, New York Israelachvili JN, McGuiggan M (1988) Forces between surfaces in liquids. Science 241:795–800 Ito T, Iwasaki Y, Narita T, Akiyoshi K, Ishihara K (2003) Controlled adhesion of human lymphocytes on electrically charged polymer surface having phosphorylcholine moiety. Sci Technol Adv Mater 4:99–104 Iwasaki Y, Akiyoshi K (2004) Design of biodegradable amphiphilic polymers: well–defined amphiphilic polyphosphates with hydrophilic graft chains via ATRP. Macromolecules 37:7637–7642 Iwasaki Y, Ishihara K (2005) Phosphorylcholine-containing polymers for biomedical applications. Anal Bioanal Chem 381:534–546 Iwasaki Y, Mikami A, Kurita K, Yui N, Ishihara K, Nakabayashi N (1997) Reduction of surfaceinduced platelet activation on phospholipid polymer. J Biomed Mater Res 36:508–515 Iwasaki Y, Nakabayashi N, Ishihara K (2001) Preservation of platelet function on 2methacryloyloxyethyl phosphorylcholine-graft polymer as compared to various watersoluble graft polymers. J Biomed Mater Res 57:72–78 Iwasaki Y, Sawada S, Nakabayashi N, Khang G, Lee HB, Ishihara K (1999) The effect of the chemical structure of the phospholipid polymer on fibronectin adsorption and fibroblast adhesion on the gradient phospholipid surface. Biomaterials 20:2185–2191 Iwasaki Y, Tabata E, Kurita K, Ishihara K (2005) Selective cell attachment to a biomimetic polymer surface through the recognition of cell-surface tags. Bioconjugate Chem 16:567–575 Iwata R, Suk-In P, Hoven VP, Takahara A, Akiyoshi K, Iwasaki Y (2004) Control of nanobiointerfaces generated from well-defined biomimetic polymer brushes for protein and cell manipulations. Biomacromolecules 5:2308–2314 Johnston EE, Bryers JD, Ratner BD (2005) Plasma deposition and surface characterization of oligoglyme, dioxane, and crown ether nonfouling films. Langmuir 21:870–881 Kadoma Y, Nakabayashi N, Masuhara E, Yamauchi J (1978) Synthesis and hemolysis test of polymer containing phophorylcholine groups. Koubunshi Ronbunshu (Jpn J Polym Sci Technol) 35:423–427 Kim K, Shin K, Kim H, Kim C, Byun Y (2004) In situ photopolymerization of a polymerizable poly(ethylene glycol)-covered phospholipid monolayer on a methacryloyl-terminated substrate. Langmuir 20:5396–5402 Kitano H, Sudo K, Ichikawa K, Ide M, Ishihara K (2000) Raman spectroscopic study on the structure of water in aqueous polyelectrolyte Solutions. J Phys Chem B 104:11425–11429 Kitano H, Imai M, Mori T, Gemmei-Ide M, Yokoyama Y, Ishihara K (2003) Structure of water in the vicinity of phospholipid analogue copolymers as studied by vibrational spectroscopy. Langmuir 19:10260–10266 Kohler AS, Parks PJ, Mooradian DL, Rao GHR, Furcht LT (1996) Platelet adhesion to novel phospholipid materials: modified phosphatidylcholine covalently immobilized to silica, polypropylene, and PTFE materials. J Biomed Mater Res 32:237–242 Kojima M, Ishihara K, Watanabe A, Nakabayashi N (1991) Interaction between phospholipids and biocompatible polymers containing a phosphorylcholine moiety. Biomaterials 12:121–124 Lee JH. Baker TJ, Mahal LK, Zabner J, Bertozzi C R, Wiemer DF, Welsh MJ (1999) Engineering novel cell surface receptors for virus-mediated gene transfer. J Biol Chem 274:21878– 21884 Lee JH, Lee HB (1993) A wettability gradient as a tool to study protein adsorption and cell adhesion on polymer surfaces. J Biomater Sci Polym Ed 4:467–481

12 Nonbiofouling Surfaces Generated from Phosphorylcholine-Bearing Polymers 325 Lee JH, Li T, Park K (2001) Solvation interactions for protein adsorption to biomaterials surfaces. In: Mora M (ed) Water in Biomaterials Surface Science. John Wiley & Sons, Chichester, pp 127–146 Lemieux GA, Yarema KJ, Jacobs CL, Bertozzi CR (1999) Exploiting differences in sialoside expression for selective targeting of MRI contrast reagents. J Am Chem Soc 121:4278– 4279 Li Y, Armes SP, Jin X, Zhu S (2003) Direct synthesis of well-defined quaternized homopolymers and diblock copolymers via ATRP in protic media. Macromolecules 36:8268–8275 Lobb EJ, Ma IY, Billingham NC, Armes SP, Lewis AL (2001) Facile synthesis of well-defined, biocompatible phosphorylcholine-based methacrylate copolymers via atom transfer radical polymerization at 20 ◦ C. J Am Chem Soc 123:7913–7914 Lu DR, Lee SJ, Park K (1991) Calculation of solvation interaction energies for protein adsorption on polymer surfaces. J Biomater Sci Polym Ed 3:127–147 Ma Y, Lobb EJ, Billingham NC, Armes SP, Lewis AL, Lloyd AW, Salvage J (2002) Synthesis of biocompatible polymers. 1. Homopolymerization of 2-methacryloyloxyethyl phosphorylcholine via ATRP in protic solvents: an optimization study. Macromolecules 35:9306–9314 Ma Y, Tang Y, Billingham NC, Armes SP, Lewis AL (2003a) Synthesis of biocompatible, stimuli-responsive, physical gels based on ABA triblock copolymers. Biomacromolecules 4:864–868 Ma Y, Tang Y, Billingham NC, Armes SP, Lewis AL, Lloyd AW, Salvage JP (2003b) Welldefined biocompatible block copolymers via atom transfer radical polymerization of 2-methacryloyloxyethyl phosphorylcholine in protic media. Macromolecules 36:3475– 3484 Mahal LK, Yarema KJ, Bertozzi CR (1997) Engineering chemical reactivity on cell surfaces through oligosaccharide biosynthesis. Science 276:1125–1128 Matsuda T, Kaneko M, Ge S (2003a) Quasi-living surface graft polymerization with phosphorylcholine group(s) at the terminal end. Biomaterials 24:4507–4515 Matsuda T, Nagase J, Ghoda A, Hirano Y, Kidoaki S, Nakayama Y (2003b) Phosphorylcholineendcapped oligomer and block co-oligomer and surface biological reactivity. Biomaterials 24:4517–4527 Marra KG, Winger TM, Hanson SR, Chaikof EL (1997) Cytomimetic biomaterials. 1. In-situ polymerization of phospholipids on an alkylated surface. Macromolecules 30:6483–6488 Oscarsson S (1997) Factors affecting protein interaction at sorbent interfaces. J Chromatography B699:117–131 Pashley RM, McGuiggan Pm, Ninham BW, Evans DF (1985) Attractive forces between uncharged hydrophobic surfaces: direct measurements in aqueous solution. Science 229:1088–1089 Patel JD, Iwasaki Y, Ishihara K, Anderson JM (2005) Phospholipid polymer surfaces reduce bacteria and leukocyte adhesion under dynamic flow conditions. J Biomed Mater Res A 73:359–366 Ratner BD (1993) New ideas in biomaterials science – a path to engineered biomaterials. J Biomed Mat Res 27:837–850 Ratner BD (1996) The engineering of biomaterials exhibiting recognition and specificity. J Mol Recognit 9:617–625 Saxon E, Bertozzi CR (2000) Cell surface engineering by a modified Staudinger reaction. Science 287:2007–2010 Singer SJ, Nicolson GL (1972) The fluid mosaic model of the structure of cell membranes. Science 175:723–731

326

Y. Iwasaki et al.

Singhvi R, Kumar A, Lopez GP, Stephanolpoulos GN, Wang DI, Whitesides GM Ingber DE (1994) Engineering cell shape and function. Science 264:696–698 Sofia SJ, Premnath V, Merrill EW (1998) Poly(ethylene oxide) grafted to silicon surfaces: grafting density and protein adsorption. Macromolecules 31:5059–5070 Sugiyama K, Mitsuno S, Shiraishi K (1997) Adsorption of protein on the surface of thermosensitive poly(methyl methacrylate) microspheres modified with the N-(2hydroxypropyl)methacrylamide and 2-(methacryloyloxy)ethyl phosphorylcholine moieties. J Polym Sci A Polym Chem 35:3349–3355 Tegoulia VA, Cooper SL (2000) Leukocyte adhesion on model surfaces under flow: effects of surface chemistry, protein adsorption, and shear rate. J Biomed Mater Res 50:291–301 Tsuruta T (1996) Contemporary topics in polymeric materials for biomedical applications. Adv Polym Sci 126:1–51 Ueda T, Oshida H, Kurita K, Ishihara K, Nakabayashi N (1992) Preparation of 2-methacryloyloxyethyl phosphory1choline copolymers with alkyl methacrylates and their blood compatibility. Polym J 24:1259–1269 Umeda T, Nakaya T, Imoto M (1982) Polymeric phospholipid analogues, 14. The convenient preparation of a vinyl monomer containing a phospholipid analogue. Macromol Chem Rapid Commun 3:457–459 Wang Y, Su TJ, Green R, Tang Y, Styrkas D, Danks TN, Bolton R, Lu JR (2000) Covalent coupling of phospholipid monolayers on the surface of ceramic materials. Chem Commun 7:587–588 Wu YJ, Timmons RB, Jen JS, Molock FE (2000) Non-fouling surfaces produced by gas phase pulsed plasma polymerization of an ultra low molecular weight ethylene oxide containing monomer. Colloids Surf B Biointerfaces 18:235–248 Yarema KJ, Mahal LK, Bruehl RE, Rodriguez EC, Bertozzi CR (1998) Metabolic delivery of ketone groups to sialic acid residues. Application to cell surface glycoform engineering. J Biol Chem 273:31168–31179 Yarema KJ, Bertozzi CR (2001) Characterizing glycosylation pathways. Genome Biol 2:REVIEWS0004

Subject Index

Acrylic acid 286, 288, 295 Adsorption kinetic constant 53, 55, 56, 58, 62 Adsorption kinetic constant (mean) 60, 68, 69 Adsorption kinetics 51, 52 Affinity constant 85, 91 AG 273 α-allyl glucoside 246, 247, 249 α-allyl glucoside (AG) 273 α-chymotrypsin 61, 64–66 α-lactalbumin 259 Amylase 291 Anchorage 195 Anhydride acid copolymers 177 Antibody 83, 84, 87 Anticortisol 84 Antiferroelectric 66 Antigen 84, 86 Anti-ovalbumin 85, 86 Asymmetry parameter 29 Atom transfer radical polymerization (ATRP) 314 Atomic force microscope 24 ATR-FTIR 119 Attenuated total reflection infrared (ATR-IR) 304 Avidin 211, 294 Bilayer 92 Binding 76 Binding calcium 91 Biomembrane 299 Biosensors 75 Biotin 88, 211, 294 Biotin streptavidin 88 Blocking 85, 86 Blood serum 46 Bovine serum albumin (BSA) 246, 273

82,

BSA 246, 274, 276, 277, 281–283 Carbon felt 227 Catalase 222, 228 Cellulosic textile 201 Channel 52, 67 Charge density 52, 64 Charged surface 2 Chitosan 259 Chymotrypsin 203 Circular dichroism (CD) 311 Coating 257 Collagen 261 Complex mixtures 46 Confocal laser scanning microscopy 181 Conformational changes 184 Conformational rearrangement 44 Convective diffusion 38 Copolymerization 246, 247, 249, 250, 267 Cotton 201, 203 Coupled plasmon waveguide resonance (CPWR) 78 Critical concentration 64 Cytochrome C 259 Damköhler number 55 Debye length 2, 66 Degree of polymerization 293 Denaturation 43 Depletion 53 Depth profile 153 Desorption time constant 182 Deuterating 94 d-gluconamidoethyl methacrylate (GAMA) 281 Differentiation 196 Diffusion boundary layer 38 Diffusion coefficient 53, 55, 58, 62 Diffusion layer thickness 55

328

Subject Index

Dimensionless waveguide parameter 29 Dipolar interactions 51 Dipole moment 66 DMAEMA 277 DP 293 DPI 75, 78, 80, 84, 95 Drug discovery 77 Dual polarisation interferometry 75 Dynamic range 97 Electric field 2, 3 Electric potential 1–3, 21 Electrocatalysis 239 Electrodialysis 223 Electron-conducting textile 227 Electrostatic interactions 308, 310 Ellipsometry 106 Endothelial cells 195 Entrapment 287 Enzyme 90, 211 Enzyme biotinylation 213 Enzyme immobilization 199, 201, 204, 286, 287, 291 Evanescent field 28, 95 Factor XIII 90 Far-field 96 Ferritin 114 Ferroelectric 66 Fibrils 195 Fibrinogen 261 Fibroblasts 249 Fibronectin 175 Fluorescence 55 Fluorophore 64 Fresnel coefficient

27

HeLa 320 HEMA 283–286

IgG3 83, 84 Immobilisation 83, 84, 86, 87 Information entropy 159 Insulin 259, 263 Interfaces 75 Interfacial reaction 57, 60 Interfacial structural transition Interfacial transition 64 Interferometric 96 Invertase 288 Ion exchanging textiles 211 Ionisation 180 Kininogen

γ -(aminopropyl) triethanoxysilane (γ -APS) 274 γ -ray 282 Glucose oxidase 252 Glycidyl methacrylate (GMA) Glycopolymer 273, 281 GMA 291 Grafting degree 253, 254 Grating coupling 34

Hemoglobin 251 Heparin 259, 261, 263 Hepatocytes 249 Heteroexchange 181 High-performance liquid chromatography 293 HL-60 320 Homoexchange 181 Horseradish peroxidase (HRP) 291 HPLC-based amino acid quantification 181 Human plasma fibrinogen 260 Human serum albumin 115, 260 Hydrogen bonding 76 Hydrophilic support 120 Hydrophilization 52 Hydrophobic interaction 301, 303 Hydrophobic supports 120 Hydrophobicity 179 2-hydroxyethyl methacrylate (HEMA) 283

286

51

258

Lévêque constant 52, 53 Lévêque constant (mean) 59 Langmuir-Blodgett technique 36 LDA 158 Lineweaver-Burk 234 Lipase 276, 286–288, 295 Liposome 92 Lysozyme 67, 81, 259 Macrophage 249, 263 Maltooligosaccharide (MOS) Mass spectra 153

291

Subject Index Maxwell’s equations 98 Membrane 214, 232, 245 Membrane mimics 92 Membrane-bound proteins 77 Memory kernel 46 Metal ion 90 2-methacryloyloxyethyl phosphorylcholine (MPC) 302 Mica 61 Michaelis-Menten kinetic 210 Microcalorimetry 78 Milk 115 Mode equation 28 Molecular recognition 212, 217 MOS 293 Multivariate analysis 158 Mutual information 159 N,N-dimethylaminorthyl methacrylate (DMAEMA) 277 N,N-dimethylformamide 273 Nanoparticles 80 Nanosphere 80 N-carboxyanhydride (NCA) 273 Neutron reflection 75, 80, 93 NH/ND isotope exchange 119 N-levulinoylmannosamine (ManLev) 319 Nonbiofouling 299 Normalized propagation constant 29 Nuclear magnetic resonance (NMR) 77 N-vinyl-2-pyrrolidone 246, 249, 250 Optical waveguide 91 Optical waveguide light spectroscopy (OWLS) 2, 6, 8, 11, 13, 55, 78 Orientation 128, 184 Ozone 285 PAN 245 PANI 291 PAP 277, 279, 287 Particle-surface interaction free energy 40 PCA 158 PEG 252 Penetration depth 29 Pepsin 201, 204, 205, 208 Peptides 76

329 Peroxidase 230 Phospholipid 263, 264, 266 Phospholipid-analogous polymers (PAP) 277, 287 Phosphorylcholine 299 Plasma 273, 276 Platelet 246, 250, 258, 259, 261, 263, 264, 266 Poly(ethylene glycol) 252 Poly(ethyleneimine) 80, 258, 259 Poly(ethyleneoxide) 258, 309 Poly(γ -ethyl-L-glutamate) 287 Poly(γ -stearyl-L-glutamate) (PSLG) 273 Poly(HEMA) 304 Poly(N-vinyl-2-pyrrolidone) (PVP) 279 Polyacrylonitrile 245 Polyaniline (PANI) 291 Polyelectrolytes 36 Polymer brushes 314 Polymer Surfaces 175 Polypropylene microporous membrane (PPMM) 273 Polypyrrole 228 Potential barrier 40 PPMM 273, 276, 277, 279, 281, 283, 285–288, 291, 295 Proteases 203 Protein adsorption 271–273, 276, 279, 281, 283, 295, 300 Protein charge 2, 19, 20, 66 Protein orientation 81, 83 Protein structure 75, 87 PSLG 274, 276 PVP 281 QCM-D 78, 188 Quartz crystal microbalance Quaternary 76 Radioactivity 61 Raman 304 Random cluster 45 Refractive index 95 Refractive index increment Reorganisation 184 Reorientation 44 Resistance 60

78

29, 31, 101

330

Subject Index

SAM 308, 312 Secondary ion image 153, 156 Secondary structures 119 Self-assembled monolayers 119 SIMS 153 Single-molecule observation techniques 24 Site-specific immobilization 287, 294 Small molecules 76 Smoluchowski 64 Solvation 119 Specific binding 76 SPR 78, 107 Stoichiometric 79 Streaming potential 51, 52, 64 Streptavidin 88 Structural transition 51 Structurally informed drug discovery 77 Structure of proteins 75 Surface modification 271–273, 276, 281, 283, 285, 286 Surface plasmon resonance 78, 94 Swelling 179 Tagless biosensor 78 Tertiary structure 76, 194 Textile 199, 211 Time-of-flight mass analyzer

153

Time-of-Flight Secondary Ion Mass Spectrometry 153 TIRE 106, 108 TIRF 55, 56, 64, 67 TOF-SIMS 153 Tosyl chloride 202, 203, 206 Total internal reflection 28, 107 Transition 65 Transport 2, 5, 13, 15–18, 55, 57, 59 Trypsin 201, 203–205, 208 Two dimensional crystal 45 Urease 213 Uricase 211, 213, 214, 219 UV 276, 277, 281 Van der Waals force Vesicle fusion 36

301

Water 303 Water content 189 Waveguide 78, 95 Xanthine oxidase 211, 213, 215, 219 X-ray 80 X-ray crystallography 75 X-ray diffraction 23, 77 Zeta potential

64, 310

Proteins at Solid-Liquid Interfaces

Proteins at Liquid Interfaces

Proteins at Liquid Interfaces (Studies in Interface Science)

Pattern Formation at Interfaces

Biopolymers at Interfaces

Amphiphiles at Interfaces

Proteins at Low Temperatures

Electrolytes at Interfaces

Colloidal Particles at Liquid Interfaces

Applied Chemistry at Protein Interfaces

Colloidal Particles at Liquid Interfaces

Biopolymers at Interfaces, Second Edition

Polymers at Surfaces and Interfaces

Colloidal particles at liquid interfaces

Proteins

Biopolymers at Interfaces, Second Edition (Surfactant Science)

Adverbs and Adverbial Adjuncts at the Interfaces

Chemical Bonding at Surfaces and Interfaces

Chemical bonding at surfaces and interfaces

Proteins

Optics and Spectroscopy at Surfaces and Interfaces

Chemical Bonding at Surfaces and Interfaces

Fluorescent Proteins

Designing Interfaces

Designing Interfaces

Designing Interfaces

Designing Interfaces

Designing Interfaces

animal proteins

Unfolded Proteins

Designing Interfaces

Proteins at Solid-Liquid Interfaces

Proteins at Liquid Interfaces

Proteins at Liquid Interfaces (Studies in Interface Science)

Pattern Formation at Interfaces

Biopolymers at Interfaces

Amphiphiles at Interfaces

Proteins at Low Temperatures

Electrolytes at Interfaces

Colloidal Particles at Liquid Interfaces

Applied Chemistry at Protein Interfaces

Colloidal Particles at Liquid Interfaces

Biopolymers at Interfaces, Second Edition

Polymers at Surfaces and Interfaces

Colloidal particles at liquid interfaces

Proteins

Biopolymers at Interfaces, Second Edition (Surfactant Science)

Adverbs and Adverbial Adjuncts at the Interfaces

Chemical Bonding at Surfaces and Interfaces

Chemical bonding at surfaces and interfaces

Proteins

Optics and Spectroscopy at Surfaces and Interfaces

Chemical Bonding at Surfaces and Interfaces

Fluorescent Proteins

Designing Interfaces

Designing Interfaces

Designing Interfaces

Designing Interfaces

Designing Interfaces

animal proteins

Unfolded Proteins

Designing Interfaces

Recommend Documents